A highly sensitive and selective signal-on strategy for microRNA quantification

Abstract

MicroRNAs (miRNAs) are associated with physiological and pathological processes. They are recognized as biomarkers for diseases diagnosis and treatment evaluation. Herein we propose a simple and cost-effective HPLC method for quantitative assay of target miRNAs with femtomolar sensitivity, single-base discrimination selectivity and low background. The assay is based on an innovative signal-on strategy. In this strategy, polyadenylation of poly(A) polymerase extends an all ‘A’ sequence at the end of target miRNA, and the substantially increased number of adenine bases are labeled with 2-Chloroacetaldehyde (CAA) to open a signal-on mode and realize a signal amplification. The linearly amplified fluorescence signal is separated from other inference signals and quantified by high performance liquid chromatography with fluorescence detection (HPLC-FD). Combining with affinity magnetic solid phase extraction (MSPE), the method is well suited for analysis of complex biological samples such as serum and cell lysate with nearly zero background fluorescence. Taking miRNA-21 as the model analyte, this absolute quantification method has a limit of detection of 200 fM and a linear calibration curve (R² = 0.999) in the range from 2.00 pM to 1.00 nM. Using locked nucleic acid (LNA) modified probes rather than ssDNA probes, the assay selectivity is improved. Moreover, analysis of bovine serum and cell lysate samples by using the method is demonstrated. Intracellular content of miRNA-21 is found to be 0.0150 amol/cell in MCF-7 cells with an assay repeatability of 4.0% (RSD, n = 3). The present HPLC quantification of miRNA offers an accurate, reliable, and cost-effective means for quantitative assay of miRNAs occurring in biological samples. Also importantly, it eliminates the need for total RNA isolation for the analysis. It may be useful for more effective diagnosis of diseases and therapeutic evaluation.

Graphical Abstract

1. Introduction

MicroRNAs (miRNAs) are non-coding, evolutionarily conserved, short single-stranded RNAs of around 22 nucleotides in length [1]. They are known to be associated with physiological and pathological processes. For example, their abnormal expression in tumors is well known, and thus, they are suggested as biomarkers for various cancers [2,3]. miRNAs have several unique characteristics, including a short length, lability to degradation, sequence homology among family members, and low abundance in total RNA samples [4]. These make quantitative assay of miRNAs in real biological samples a big challenge. Therefore, developing highly sensitive and selective methods for miRNA assay is in demand.

Current methods for miRNA detection include Northern blotting [5] and microarrays [6]. They are considered in many cases to lack assay sensitivity and repeatability. It’s well known that the sensitivity can be improved by nucleic acid amplification, e.g. quantitative reverse transcription-polymerase chain reaction (RT-qPCR) [7]. RT-qPCR is sensitive, but it requires different temperature adjustment procedures and sophisticated probe designs, which significantly increase the experimental complexity and test costs. Recently, isothermal amplification-based assays involving various amplification schemes such as rolling circle amplification [8,9], hybridization chain reaction [10], catalytic hairpin assembly [11], loop-mediated isothermal amplification [12], strand-displacement amplification [13,14], duplex-specific nuclease signal amplification [15], and nanomaterials based amplification [16,17] have been developed. Most of these methods deploy fluorescence or electrochemical detection. They normally require total RNA isolation to eliminate interference from sample matrix and using fluorescently labeled or electrochemically active probes. In addition, most of these assays are compromised by a high background signal [8,13–17].

High performance liquid chromatography with fluorescence detection (HPLC-FD), a widely accessible analytical technique in the lab, offers advantages such as high sensitivity, accuracy, reproducibility, and ease of automation. While HPLC-FD is among the most popular techniques used in bioanalysis, quantification of miRNAs by using an HPLC based method hasn’t been reported so far. In this work, we aimed to develop a highly sensitive and selective method for quantitative assay of miRNA deploying HPLC-FD. We proposed a signal-on strategy. Locked nucleic acid (LNA)-modified DNA probes were immobilized on magnetic beads and tested for isolation of the targeted miRNA from sample matrix via selective hybridization and subsequent magnetic separation. It has been demonstrated that LNAs modified probes have a superior hybridization affinity over ssDNA probes towards its complementary nucleotide sequence [18,19]. Moreover, LNA modified probes display a high nuclease resistance, which makes them better suited for uses in complex sample matrices such as serum [20,21]. To achieve high assay sensitivity, signal amplification based on miRNA polyadenylation by poly(A) polymerase was investigated. Hydrolysis of the extended target miRNA by formic acid and HPLC-FD determination of free adenine, the reporter for target miRNA were studied. In this work, miRNA-21 was taken as the model analyte. Previous studies have shown that miRNA-21 is up-regulated in serums linked to ovarian cancer, diffused-typed B cell lymphoma, and breast cancer prostate cancer [22–25]. Quantification of miRNA-21 in biological samples, i.e. bovine serum and cell lysates by using the method proposed was finally demonstrated.

2. Methods

2.1. Chemicals and reagents

Chloroacetaldehyde dimethyl acetal, hydroxymethyl amino-methane (Tris), sodium chloride (NaCl), hydrochloric acid (HCl), formic acid, adenine, bovine serum, triethylamine (TEA), acetic acid, ethylenediaminetetraacetic acid (EDTA), Triton X-100, and concentrated H₂SO₄ were obtained from Sigma-Aldrich. SYTO™ RNASelect™ green fluorescent cell stain 5 mM solution was purchased from ThermoFisher Scientific. PBS tablets were from Gibco, 5′-UAG CUU AUC AGA CUG AUG UUGA-3’ (miRNA-21), 5′-TCA GTC TGA TAA GCT A -TEG-biotin-3’ (LNA modified probe, all LNA bases are underlined) was purchase from Exiqon. 5′-UAG CUU AUC GGA CUG AUG UUGA-3′ (single-based mismatch), 5′- UUG CUU AUC GGA CUG AUC UUGA-3’ (three-based mismatch), and 5′-GUA AGG CAU CUG ACC GAA GGCA-3’ (non-complementary) (all HPLC purified grade) were customer ordered from Sigma-Aldrich. Pierce™ Streptavidin magnetic beads (1 μm) and diethylpyrocarbonate (DEPC) treated water were from Invitrogen. E. coli poly(A) polymerase and RNase inhibitor were from New England Biolabs Ltd. All miRNA stock solutions were prepared daily with DEPC-treated water in an RNase-free environment. To avoid the effect of RNase on the stability of miRNAs, all centrifuge tubes and tips and buffer prepared were autoclaved, and DEPC-treated water was used throughout this work. Binding/wash buffer was composed of 10 mM Tris buffer at pH 7.5, 2 M NaCl, 1 mM EDTA, and 0.0005% (w/v) Triton X-100.

2.2. Instrumentation

A Shimadzu 20A HPLC system (Shimadzu, Nakagyo-Ku, Kyoto, Japan) that included a DGU-20A5 degasser, a LC-20 AD liquid chromatograph, a LC-20AHT auto sampler, an RF-10AXL fluorescence detector, and a CTO-20A column oven was used. LC Solutions Analyst software from Shimadzu was used for data acquisition and analysis. A Pinnacle II C₁₈ column (200 × 4.60 mm, particle size 5 μm) from RESTEK was used for HPLC separation.

2.3. Preparation of 2-chloroacetaldehyde (CAA)

The procedure previously reported [26] was followed with modifications. Briefly, equal volumes of chloroacetaldehyde dimethyl acetal and diluted H₂SO₄ (1 + 9) were mixed and then distilled slowly under a fume hood. The distillate fraction containing CAA (at ca. 1.5 M) was collected between 80 and 85 °C and stored in darkness at 4 °C before use.

2.4. Preparation of LNA modified probe-magnetic bead conjugate

Streptavidin-magnetic bead suspension (50 μL) was transferred into to vial. The beads were washed twice by the binding/washing buffer, and then re-suspended in 100 μL of the buffer. Biotin-LNA modified probe (5 × 10⁻⁴ mol/L, 5 μL) was added. The mixture was placed on a shaker at 37 °C for 30 min. Magnetic beads were then collected and washed three times with the binding/washing buffer to remove any unbound biotin-LNA modified probe. The conjugate collected was re-suspended in 1.0 mL binding/washing buffer.

2.5. HPLC-FD assay of miRNA-21

miRNA-21 containing sample solution (10 μL) was added to LNA modified probe-magnetic bead conjugate suspension (200 μL). The mixture was incubated at 37 °C for 30 min to complete the hybridization. After magnetic separation, the supernatant was removed and 1 × poly(A) buffer (50 mM Tris-HCl, 250 mM NaCl, 10 mM MgCl₂, 0.0005% Triton-X 100, pH 7.9), ATP (10⁻⁶ M), poly(A) polymerase (0.05 U/μL), and RNase inhibitor (0.2 U/μL) were added to a final volume of 100 μL and incubated at 37 °C for 2 h for polyadenylation. The magnetic beads were collected and washed twice with the binding/washing buffer to remove any unbounded species such as ATP. The beads were re-suspended in 10 μL of the binding/washing buffer. Heteroduplexes immobilized on magnetic beads were denatured at 90 °C for 3 min, releasing extended miRNA-21 into solution. After removal of the magnetic beads, 80% (v/v) formic acid (90 μL) was added. The solution was heated at 150 °C for 10 min to hydrolyze ploy(A) tagged miRNA-21. The hydrolysis solution was evaporated at 40 °C for 10 min using a vacuum concentrator. 1 × PBS (90 μL) and CAA solution (10 μL) were added to the residue, then the solution was heated under a fume hood at 100 °C (in boiling water) for 5 min. After heating, the reaction solutions were placed on an ice bath to cool down. Portions (10 μL) of the solution were injected into the HPLC system for analysis. Separation was performed under isocratic conditions at a flow rate of 1.0 ml/min with a mobile phase consisting of 0.1 M triethylammonium acetate (TEAA) buffer at pH 7.0 and acetonitrile (93:7 v/v). Column temperature was maintained at 30 °C. For fluorescence detection, the excitation and emission wavelengths were set at 490 and 530 nm, respectively.

3. Results and discussion

3.1. Assay design

To achieve quantitative assay of miRNAs occurring in biological samples, one faces challenges including (but not limited to) selective and quantitative isolation of the targeted miRNAs, and sensitive and repeatable determination of them. In this work, a novel signal-on assay strategy based on a combination of affinity magnetic solid phase extraction (MSPE), enzymatic isothermal amplification, and HPLC-FD is developed to overcome the difficulties encountered in miRNA quantification. As shown in Scheme 1, the experimental workflow consists of 3 steps: (1) isolation of target miRNA using LNA modified probe-magnetic bead conjugate; (2) polyadenylation of target miRNA by poly(A) polymerase; and (3) HPLC-FD determination of ε-adenine after acid hydrolysis of the extended miRNA target and CAA labeling.

Scheme 1.

Quantitative assay of miRNAs by HPLC-FD based on a signal-on strategy.

In this work, LNA modified probes were deployed for select hybridization with the targeted miRNAs. It has been shown that LNA modification offers various advantages, including improved affinity towards complementary miRNA and better stability of the probes [21–24]. It is important that the probes are designed to be ~6-mers shorter than target miRNA so that the heteroduplex formed has an overhang six bases sequence at the 3′-OH end. This particular structure is easily recognized by poly(A) polymerase, ensuring a high efficacy of polyadenylation reaction. To isolate target miRNAs from sample matrix, LNA modified probes were conjugated with magnetic beads through biotin and streptavidin reaction. Target miRNA was captured on magnetic beads through select hybridization and isolated from sample matrix by convenient magnetic separation. A poly(A) tail was created at the 3′-OH end of the bound target miRNA via polyadenylation reaction. After removal of the magnetic beads, the extended miRNA was released from magnetic beads into solution by denaturing the heteroduplex. It’s worth noting that a poly(A) tail stabilizes single-strand RNA by protecting it from enzymatic degradation [27]. The extended miRNA was then hydrolyzed by formic acid, resulting in a large number of free adenine (i.e. the reporter for the targeted miRNA) in solution. Formic acid is known to degrade nucleic acids into nucleobases [28]. Compared with enzyme-catalyzed hydrolysis, acid hydrolysis offers advantages such as high efficiency, simple experimental implementation, low test costs, and less prone to interference from test conditions such as enzyme inhibition [29–32]. To determine adenine sensitively and selectively, HPLC-FD was deployed after pre-column derivatization. CAA was used as the derivatization reagent owing to its ability to effectively label nucleobases with a broad linear range (10⁻⁹e10⁻⁴ M) and non-fluorescent property [33–35]. Although CAA itself was non-fluorescent, many fluorescent species such as etheno-adenine (ε-adenine), etheno-guanine (ε-guanine), etheno-cytosine (ε-cyto-sine), and CAA degradation products were formed under the reaction conditions. Therefore, HPLC was deployed to separate ε-adenine from all other fluorescent species present in the derivatization solution, achieving a highly sensitive fluorescence detection with nearly zero background fluorescence.

3.2. Investigation of miRNA polyadenylation on the surface of magnetic beads

To assess the feasibility of the assay designed as above, a fluorescence spectroscopic study was performed to see whether target miRNA could be effectively captured by the prepared LNA modified probe-magnetic bead conjugates and then a poly(A) tail be added to it through polyadenylation by poly(A) polymerase. An RNA dye that distinguished between short and long RNA (SYTO™ RNASelect™ green fluorescent cell stain) was used for the study. The non-fluorescent dye shows bright green fluorescence when bound to RNA (Ex/Em ~490/530 nm). The dye was added to three solutions of different compositions: #1) a poly(A) polymerase solution; #2) a miRNA-21 solution obtained from denaturation of the heteroduplex formed by miRNA-21 and LNA modified probe-magnetic bead conjugates; and #3) an extended miRNA-21 solution obtained from denaturation of the heteroduplex formed by miRNA-21 and LNA modified probe -magnetic bead conjugates after polyadenylation by poly(A) polymerase. The fluorescence data obtained are shown in Fig. 1. As shown, no fluorescence was observed from the poly(A) polymerase solution (Fig. 1, curve a), which means that poly(A) polymerase is not stained by the dye and therefore, its presence does not interfere with the fluorescence spectroscopic study. From Fig. 1 curve b, incubation of miRNA-21 with the RNA dye produced no observable fluorescence signal likely because of its very short sequence. However, after miRNA-21 polyadenylation by poly(A) polymerase, the extended miRNA-21 was stained by the dye and strong fluorescence was observed from the incubation solution (Fig. 1, curve c). This is because that extended miRNA-21 has a complex folding structure, and thus, binds dye molecules more effectively, leading to an increase of fluorescence intensity. These results indicate that target miRNA-21 was effectively captured by the LNA modified probe-magnetic bead conjugates from a solution and extended by poly(A) polymerase on the surface of magnetic beads. It’s worth noting that the reaction solution of polyadenylation by poly(A) polymerase contained Triton X-100 that was used in the washing/binding buffer to prevent magnetic beads from adsorbing on the surface of test tubes. It was, therefore, studied whether the presence of Triton X-100 affected the activity of poly(A) polymerase. Fluorescence intensity was measured for two incubation solutions with (at 0.0005% w/v) or without Triton X-100. It was found that for both solutions the fluorescence intensity was the same (Fig.S1). The presence of Triton X-100 didn’t affect polyadenylation of miRNA-21 by poly(A) polymerase.

Fig. 1.

A fluorescence spectroscopic study of polyadenylation of miRNA-21 on the surface of magnetic beads: fluorescence spectrum of a poly(A) polymerase solution (a); a miRNA-21 solution (b); and an extended miRNA-21 solution (c), indicating that the isolation of miRNA-21 by affinity MSPE, signal amplification by polyadenylation, and release of extended miRNA-21 by denaturation were successful.

3.3. Quantification of ε-adenine by HPLC-FD

ε-Adenine is designated as the signal reporter for the targeted miRNA in the assay. Its concentration is converted to that of target miRNA. Therefore, determination of ε-adenine with selectivity and repeatability is critical. Potential interference with the determination comes from the other fluorescent products formed during the pre-column derivatization, including ε-cytosine, ε-guanine, and CAA degradation products. ε-Cytosine has a very low fluorescence quantum yield. More fluorescent ε-adenine and ε-guanine have the same excitation and emission wavelength (Fig.S2). It was found that the three etheno nucleobases could be base-line separated on a C₁₈ column by using a TEAA modified water/acetonitrile mobile phase. Fig. 2 shows a typical chromatogram obtained from separating a derivatization solution containing adenine, guanine, and cytosine. As shown, ε-cytosine, ε-guanine, and ε-adenine were eluted out at 2.85, 4.50, and 6.65 min, respectively. By comparing the fluorescence intensity, fluorescence yield ratio of ε-cytosine: ε-guanine: ε-adenine is 1:87: 870, i.e. ε-adenine possesses the highest fluorescence yield among the three. An HPLC effluent fraction was collected at 6.65min to further identify the eluted compound by ESI-MS/MS. MS spectrum obtained is shown in Fig. 3. The identity of ε-adenine is validated by the molecular ion m/z 160.03 and its three typical fragment ions m/z 133.13, m/z 106.11 and m/z 119.05.

Fig. 2.

HPLC separation of a mixture of adenine (0.5 μM), guanine (5 μM), and cytosine (95 μM) after pre-column derivatization with CAA.

Fig. 3.

MS² spectrum of ion m/z 160 in an HPLC effluent fraction collected at 6.65 min, validating the identity of ε-adenine as the eluted compound.

3.4. Study of experimental conditions

Incubation time for polyadenylation affects the prolongation of miRNA by poly(A) polymerase. Heteroduplex of miRNA-21 and LNA modified probe-magnetic bead conjugate was prepared and washed twice with the binding buffer. It was then incubated with poly(A) polymerase and ATP. Portions of the incubation solution were taken at different times and denatured. After acid hydrolysis of the extended miRNA-21, ε-adenine was determined. The results are illustrated in Fig. 4A. As shown, length of the poly(A) tail (i.e. ε-adenine concentration in the hydrolysis solution) increased as incubation time increased until ~120 min when the maximum was reached. An incubation time of 120 min was selected for further study.

Fig. 4.

Study of experimental conditions: effects of incubation time for polyadenylation (A), and reaction time for fluorescence tagging of adenine with CAA (B). Error bars represent the standard deviation of three independent tests.

In previous reports on labeling of nucleic acids and nucleobases with CAA, the labeling reaction was performed at either room temperature or at 80 °C. The labeling took hours to complete. In this work, it was found that adenine could be effectively labeled with CAA at 100 °C, producing highly florescent ε-adenine. Under the derivatization conditions, ε-adenine concentration quickly reached a maximum in less than 5 min and remained almost the same till 10 min as shown in Fig. 4B. The results indicate that ε-adenine formed was very stable even at an elevated temperature. This property facilitates its subsequent determination by HPLC-FD. For further study, 5 min was chosen as the reaction time for pre-column derivatization. Acid hydrolysis of extended target miRNAs was involved in the assay. Effects of hydrolysis time and concentration of formic acid were investigated (Fig.S3). The hydrolysis was found complete at ca. 10 min. and the hydrolysate remained stable till about 60 min. For further tests hydrolysis for 10 min was performed. Various concentrations of formic acid, i.e. 20%, 40%, 60%, 80%, 98% (v/v) were tested. It was found that using formic acid at 60%, 80%, and 98% (v/) produced similar results. Therefore, a concentration of formic acid was chosen at 80% (v/v) for the hydrolysis.

3.5. Quantitative assay of miRNA-21

Under the experimental conditions selected, analytical figures of merit were studied, taking miRNA-21 as the model analyte. A set of water solutions containing miRNA-21 at various concentrations ranging from 2.0 pM to 1.00 nM were analyzed. Triplicate analyses were carried out for each sample. Fig.S4 shows HPLC-FD chromatograms obtained from analysis of these standard solutions. It looks like only one peak at 6.65 min appearing in the chromatogram. This is because the y-axis scale is very big. Actually, in addition to ε-adenine, several fluorescent compounds were eluted out, and their fluorescence signals became very significant when miRNA-21 concentration was low as shown in the chromatogram obtained from analysis of a serum sample spiked with miRNA-21 at 10 pM (Fig. 5). It’s also worth mentioning that other big peaks were detected at cleaning up the column with a high acetonitrile-content mobile phase. Linear regression analysis on the fluorescence intensity/miRNA-21 concentration data yielded linear calibration curves for two concentration ranges, respectively:

$$\text{FI}=956.85\text{C}+484.2R^2=0.99932.0\text{\hspace{0.17em}pM}-80.0\text{\hspace{0.17em}pM}$$

$$\text{FI}=949.89\text{C}+7272.7R^2=0.999680.0\text{\hspace{0.17em}pM}-1.0\text{\hspace{0.17em}nM}$$

where FI was fluorescence intensity in μV and C was miRNA-21 concentration in pM. It’s noted that the two linear regression lines are different only in their intercepts. The slopes are the same. It’s worth mentioning that a calibration curve covering the whole concentration range tested (i.e. 2.0 pM to 1.0 nM) was also linear with r² of 0.9997. However, in order to reflect the reasonable weights for data at low concentrations (2.0pMe80.0pM), we chose to split it into two ranges. These calibration curves were used accordingly based on the concentration of analytes in the sample under investigation. The limit of detection was estimated to be 200 fM (S/N = 3). This remarkably low limit of detection can be ascribed to two factors: 1) a nearly zero background signal and 2) linear amplification of the fluorescence reporter. A high fluorescence background usually leads to a high limit of detection as it’s calculated from the variation of the background signal. The comparisons of related references were listed in Table.S1. Another significant advantage of the present assay is that the fluorescence signal is linearly proportional to miRNA concentration. In most fluorometric assays involving endo-/exo-nuclease based isothermal amplification the fluorescence signal is linearly proportional to the logarithm of miRNA concentration [14,36–39]. With a logarithmic calibration curve, a 10-fold increase in miRNA concentration merely doubles the fluorescence signal. However, with a linear calibration curve of fluorescence signal versus miRNA concentration, the increase in fluorescence signal is 10 times, which obviously allows measurements with far better accuracy and repeatability. To assess the assay repeatability, a miRNA-21 solution at 5.0 pM was analyzed for 5 times. Peak intensity and the retention time of ε-adenine were recorded. Repeatability (RSD) of peak intensity was calculated to be 1.8%, and that of retention time was 1.3% (n = 5). These results indicate the proposed HPLC-FD method offers a very good assay repeatability. In comparison with the results in the literature, very recently, several groups have separately developed fluorescence quantitative assays of miRNA on the basis of a 1:1 binding mode, namely, one target molecule can only hybridize with one fluorescent DNA probe and obtain the fluorescence of single fluorophore. The limit of detection of these assays for miRNA is generally in the range of 100 pM to 10 nM [40–43], indicating that the sensitivity of the demonstrated assay is more than 3 orders of magnitude higher. Compared with these fluorescent miRNA assays of 1:1 binding mode, in this proposed assay, one target miRNA can produce a long all adenine tail (estimated to be ca. 250 adenine bases), therefore, it’ll generate numerous fluorophores after CAA labeling and then significantly increase the fluorescence signal, resulting in the greatly improved sensitivity. Moreover, unlike endo-/exonuclease, poly (A) polymerase used in this paper does not require specific site recognition and probe design, which is another advantage. Comparing this proposed method to recent literatures about miRNA quantification based on poly (A) polymerase amplification, it is more than 3 orders of magnitude higher than that of SYBR Green I (0.53 nM) [44], and is comparable to the complicated microarray detection [45,46].

Fig. 5.

HPLC-FD chromatogram obtained from analysis of a serum sample spiked with miRNA-21 at 10.0 pM.

3.6. Selectivity of the assay

To evaluate the selectivity of the proposed assay, standard solutions (at 80.0 pM) of miRNA-21 (5′-UAG CUU AUC AGA CUG AUG UUGA-3′), single-based mismatched miRNA (5′-UAG CUU AUC GGA CUG AUG UUGA-3′), three-based mismatched miRNA (5′-UUG CUU AUC GGA CUG AUC UUGA-3′), and non-complementary miRNA (5′-GUA AGG CAU CUG ACC GAA GGCA-3′) were analyzed. The results of peak intensity for ε-adenine are summarized in Fig. 6. As shown, the fluorescence signal obtained from non-complementary microRNA was almost the same as that from a blank, i.e. nearly zero. The fluorescence responses obtained from the single-base mismatched and three-base mismatched miRNA solutions were 4.1% and 2.2% of that obtained from the miRNA-21 solution, respectively. Based on these results, a single base mismatched miRNA at the same concentration of target miRNA will cause a positive error of 4.1% in the assay results. It’s worth noting that when a ssDNA probe is used in the assay the error is ca. +10% (Fig.S5). The results prove the present assay offers an improved selectivity by deploying LNA modified probes.

Fig. 6.

Specificity of the assay proposed for miRNA quantification. Error bars represent the standard deviation of three independent experiments.

3.7. Quantification of miRNA-21 in serum and cell lysate samples

Analysis of clinical samples such as serum normally involves a tedious procedure for sample clean-up. In miRNA assays, total RNA isolation is usually performed. As summarized in Table.S1, the majority of isothermal amplification fluorometric assays previously reported were shown useful for analysis of total RNAs. However, miRNAs are easily lost during the procedure of total RNA isolation by virtue of their short sequences and lability to degradation. To access the potential of the present assay in biological and clinical analysis, it was applied to determine miRNA-21 in bovine serum and cell lysates. A bovine serum sample was thawed and diluted with the binding buffer (1 + 9). RNase inhibitor was added. Five replicate runs were conducted on 100 μL portions of the diluted serum sample. The analytical data are summarized in Table 1. As shown, no miRNA-21 was detected in this bovine serum sample. The sample was spiked with authentic miRNA-21 at 5, 50, and 500 pM, and the recovery was found in the range of 98.4–102.6% with a good assay repeatability (RSD <4.2% in all three cases). These results indicate that there was no interference with the quantification from the endogenous compounds in this sample matrix. Determination of intracellular miRNA-21 by using the present assay was also demonstrated. MCF-7 cells were collected, washed with PBS twice, and lysed in the binding buffer with a cell disruptor. Three cell lysates prepared in different days were analyzed. miRNA-21 was found to be 15.0 pM in the cell lysate (10⁶ cells/mL lysate) or0.0150 amol/cell (i.e. 9030 copies/cell), which was consistent with previously reported results [47–50]. To further assess the assay accuracy, miRNA-21 was spiked into the cell lysate samples, and the samples were analyzed again. The results are also listed in Table 1. Recovery of miRNA-21 from this cellular sample was found in the range from 98.6 to 103.3%. These results indicate that the proposed assay is accurate, repeatable, and without needing any sample pretreatments such as total RNA isolation. Therefore, it has a good potential in analysis of biological samples.

Table 1.

Determination of miRNA-21 recovery in bovine serum samples and cell lysate samples.

	Added (pM)	Found (pM)	Recovery (%)	RSD (%, n = 3)
Bovine serum	0	ND	_	_
	5.0	4.92	98.4	4.2
	50.0	51.3	102.6	3.0
	500.0	496.2	99.2	3.2
	Added (amole)	Found (amole)	Recovery (%)	RSD (%, n = 3)
MCF-7 cell line (1 × 104 cells)	0	150.0	_	4.0
	50.0	197.0	98.6	4.6
	500.0	670.0	103.3	2.8
	5000.0	5100.0	99.0	3.5

4. Conclusions

The first HPLC method for quantitative assay of target miRNAs has been developed in this work. The method is based on a combination of affinity magnetic solid phase extraction using LNA modified probes, isothermal signal amplification by poly(A) polymerase, and HPLC determination of the miRNA reporter with fluorescence detection. The easily handled assay is sensitive enough to detect miRNA in real sample. Other advantages include a nearly zero fluorescence background, a linear calibration curve between fluorescence signal and miRNA concentration (instead of logarithm of miRNA concentration), and improved assay selectivity due to the use of LNA modified probes (versus ssDNA probes). Quantification of miRNA-21 in bovine serum and MCF-7 cell lysates by using the present method was demonstrated. Without total isolation of RNAs prior to the assay potential loss of miRNAs was avoided. The analytical results obtained were accurate and repeatable. Considering the features of high assay sensitivity, specificity and simple implement, this method is very promising to offer a cost-effective and reliable means for quantitative assay of target miRNAs occurring in biological samples.

HIGHLIGHT.

• An HPLC method for miRNA quantification with femtomolar sensitivity and single-base discrimination selectivity is reported.

• Advantages of the proposed method include a low fluorescence background and linear signalconcentration calibration curves.

• This is the first HPLC assay for quantifying microRNAs based on isothermal signal amplification.

• Real-life samples including serum and cell culture were analyzed without any sample pretreatments.

Abbreviations

HPLC

high performance liquid chromatography

FD

fluorescence Detection

MSPE

magnetic solid phase extraction

miRNA

microRNA

MS

mass spectrum

CAA

2-chloroacetaldehyde

MW

molecular weight

PBS

phosphate buffered saline

ACN

acetonitrile

TEAA

triethylamine acetate

Conc

concentration

FI

fluorescence intensity