• Open Access

A novel real-time polymerase chain reaction method for high throughput quantification of small regulatory RNAs


*(fax +1 314 694 3914; e-mail heping.yang@monsanto.com)


MicroRNAs (miRNAs) and small interfering RNAs (siRNAs) are important players of both transcriptional and post-transcriptional gene silencing networks. In order to investigate the functions of these small regulatory RNAs, a system with high sensitivity and specificity is desperately needed to quantitatively detect their expression levels in cells and tissues. However, their short length of 19–24 nucleotides and strong similarity between related species render most conventional expression analysis methods ineffective. Here we describe a novel primer for small RNA-specific reverse transcription and a new TaqMan technology-based real-time method for quantification of small RNAs. This method is capable of quantifying miRNA and siRNA in the femtomolar range, which is equivalent to ten copies per cell or fewer. The assay has a high dynamic range and provides linear readout of miRNA concentrations that span seven orders of magnitude and allows us to discriminate small RNAs that differ by as little as one nucleotide. Using the new method, we investigated the expression pattern of gma-miRMON1, a novel miRNA identified from soybean leaves. The results were consistent with our results obtained from Northern blot analysis of gma-miRMON1 and Affymetrix microarray analysis of the gma-miRMON1 precursor, suggesting that the new method can be used in transcription profiling.


Multicellular organisms depend on spatially and temporally coordinated complex networks of gene regulatory pathways. MicroRNAs (miRNAs) are key components of these networks. An miRNA is complementary to a part of one or more mRNAs, usually at a site in the 3′UTR. The annealing of miRNA to mRNA inhibits protein translation and/or triggers the degradation of the mRNA transcript through a process similar to RNA interference (RNAi). miRNAs were initially discovered in Caenorhabditis elegans (Lee et al., 1993; Reinhart et al., 2000). To date, over 900 miRNAs have been found in various plants and animals as well (Jones-Rhoades et al., 2006). Mature miRNAs are typically ~21 nucleotides in length and are cleaved from their larger stem-looped precursors. miRNAs play a variety of fundamental regulatory roles in growth, development, differentiation, metabolism, apoptosis and cell death, and responses to biotic and abiotic stresses across eukaryotic species (Sunkar and Zhu, 2004; Sunkar et al., 2007). Abnormal miRNA levels are observed in many human diseases, including cancer; this suggests that miRNAs are legitimate targets for anti-cancer therapy (Caldas and Brenton, 2005).

RNAi, a phenomenon first observed in plants (Hamilton and Baulcombe, 1999) and now recognized as a widespread, perhaps universal mechanism, is triggered by small interfering RNAs (siRNAs). These siRNAs, 21–25 nucleotides long, are derived from the processing of the trigger double-stranded RNA and guide the RNA-Induced Silencing Complex to the target mRNA in a sequence-specific manner to mediate catalytic cleavage. The double-stranded siRNAs are used to prevent the transcription of viral RNA in nature and to most efficiently suppress gene expression for functional genomics researches and trait/drug developments. For example, the RNAi pathway can be exploited to control insect pests via in planta expression of a double-stranded RNA (Baum et al., 2007).

Northern blotting has often been used to validate and quantify miRNA expression (Jonstrup et al., 2006). A major drawback of this method is its poor sensitivity, especially when monitoring expression of low-abundant miRNAs (Table 1). Different efforts have been made to increase sensitivity and specificity of Northern blotting for small RNA detection, such as developing the liquid hybridization protocol (Overhoff et al., 2004), using oligonucleotide probes modified by locked nucleic acid (LNA) (Válóczi et al., 2004), digoxigenin-labelled RNA probes (Ramkissoon et al., 2006) or padlock probes (Jonstrup et al., 2006), and detecting confocal laser-induced fluorescence (Neely et al., 2006). However, a relatively large amount of RNA per sample is still required, which is not feasible when the cell or tissue source is limited.

Table 1.   Comparisons of small RNA quantification methods
MethodNorthern blottingInvader assayRNA-primed, array-based Klenow enzyme assayMicroarraySYBR-based PCRTaqMan PCR
RT primerNoNoNoNoPoly(T) adapterGene-specific primer with a tailStem-loop primerKey-like primer
Probe32P-labelled oligoFluoresce resonance energy transfer oligoOilgo complementary to miRNAOilgo complementary to miRNANoNoFluoresce resonance energy transfer oligoFluoresce resonance energy transfer oligo
InstrumentFilm developer/imagerFluorescence plate readerPrinter, hybridization apparatus and laser scannerPrinter, hybridization apparatus and laser scannerCyclerCyclerCyclerCycler
User bench skillModerateLowHighHighLowLowLowLow
Sensitivity>500 ng total RNA50–100 ng total RNA<10 pg miRNA>50 ng total RNA>100 pg total RNA>10 pg total RNA>25 pg total RNA>10 pg total RNA
Specificity4 mismatches1 mismatch3 mismatches2 mismatches1 mismatches1 mismatch1 mismatch1 mismatch
Dynamic range2 orders3–4 logs3 orders2 orders6 order6–7 order7 orders7 orders
ReferenceOverhoff et al., 2004 Válóczi et al., 2004Allawi et al., 2004Nelson et al., 2004Thomson et al., 2004 Shingara et al., 2005Shi and Chiang, 2005Raymond et al., 2006Chen et al., 2005This study

As the number of identified miRNA genes increased, oligoarray-based technologies to investigate the expression of hundreds of miRNAs at the same time have been developed (Babak et al., 2004; Barad et al., 2004; Liang et al., 2005; Miska et al., 2004; Nelson et al., 2004; Sempere et al., 2004; Shingara et al., 2005; Sun et al., 2004; Thomson et al., 2004). Microarray technology has several advantages over Northern hybridization, which include higher throughput, better normalization and improved sensitivity (Table 1). However, both these methods rely on an advanced read-out system and their level of specificity is not always addressed. Moreover, the short nature of the mature miRNA makes it impossible to optimize probes for all miRNAs, which therefore limits the ability to discriminate highly similar miRNAs. As a result, non-identical but closely related miRNAs are not easily distinguishable by the current microarray methods.

The Invader miRNA assay based on fluorescence detection has the ability to detect as few as 20 000 molecules of an individual miRNA in as little as 50–100 ng of total cellular RNA or as few as 1000 lysed cells (Table 1). It distinguishes between miRNAs and their precursors, as well as between closely related miRNA isotypes (Allawi et al., 2004). However, some important, biologically active endogenous or introduced miRNAs or siRNAs may be present at levels that are too low to be detected or quantified directly with technologies (Table 1) such as cloning (Sunkar et al., 2005), Northern hybridization, microarray analysis and Invader assay (Kim et al., 2004; Sempere et al., 2004).

Real-time polymerase chain reaction (PCR) is a much more sensitive method (Table 1); however, the short length of miRNAs and siRNAs presents a unique challenge in PCR design. Most conventional PCR primers are similar in length to miRNAs and siRNAs, implying that a reverse transcription primer that is longer than miRNA/siRNA would be required in order to design both forward and reverse primers. Here we describe a novel primer for small RNA-specific reverse transcription and a new design for the TaqMan-based real-time PCR. The reverse transcription primers consist of a common sequence on the 5′ end and a small RNA-specific sequence on the 3′ end. The TaqMan probe including three nucleotides identical to the 3′ end of a miRNA/siRNA is complementary to the reverse transcription primer, and the reporter dye is attached to the miRNA/siRNA nucleotide for better specificity.


Novel design of primers for reverse transcription from small RNAs

Small RNAs (miRNA and siRNA) play very important regulatory roles in transcriptional and post-transcriptional gene silencing. Quantification of these small RNAs in cells and tissues has become increasingly important. Northern blotting, which has been often used, is time-consuming. Moreover, low-abundant miRNAs have been difficult to detect based on either Northern hybridization or other available technologies, such as cloning, Invader assay and microarray. In order to develop a TaqMan technology-based, high-throughput and sensitive method for accurate quantification of small RNAs, we designed a reverse transcription primer that allows efficient and specific synthesis of cDNA from an individual small RNA. Each primer consists of two parts: on the 5′ end is a universal sequence of 35 nucleotides (5′-CGCGAGCACAGAATTAAT ACGACTCACTATACGCG-3′) and on the 3′ end 8 nucleotides complementary to a particular small RNA. The universal sequence contains 48.6% of GC. It was designed not to share homology with sequences in the public databases. The rest was to ensure small RNA-specific reverse transcription. The reverse transcription primers (5′-CGCGAGCACAGAATT AATACGACTCACTATACGCGNNNNNNNN-3′) were designed to form a partially double-stranded key-like structure that can increase reverse transcription efficiency while limit non-specific annealing (Figure 1). Each of these reverse primers play three functions: (i) to convert the RNA template into cDNA; (ii) to introduce a ‘universal’ PCR binding site to one end of the cDNA molecule; and (iii) to extend the length of the cDNA to facilitate subsequent monitoring by quantitative PCR. Previously, a stem-loop primer (5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACNNNNNN-3′) for miRNA quantification was reported (Chen et al., 2005). Although these two primers take a similar secondary structure, their sequences are completely different. The stem-loop primer is 44 nucleotides long while the key-like primer is only 35 nucleotides long. The stem-loop primer has a much longer double-stranded stem (14 pairs including both A=T and G=C) while the key-like primer has just 4 pairs and they all are G=C. The stem-loop primer contains 56.8% GC while the key-like primer has 48.6%.

Figure 1.

 Assay design of TaqMan real-time PCR for small RNA. TaqMan-based real-time quantification of small RNAs includes two steps: cDNA synthesis and real-time quantitative PCR. Small RNA-specific reverse transcription primers bind to the 3′ portion of small RNA molecules and are reverse transcribed with reverse transcriptase. cDNA product is then PCR-quantified using small RNA-specific forward primers and dye-labelled TaqMan probes with the universal reverse primer.

Sensitivity, dynamic range and specificity of the TaqMan-based real-time PCR method

We used the two-step RT-PCR for small RNA quantification. In the first step, a gene-specific key-like primer was used to specifically synthesize cDNA from a small RNA. In the second step, the reverse transcription products were quantified by conventional TaqMan real-time PCR using a combination of a miRNA/siRNA-specific forward primer and a generic universal primer common to all assays. To validate the assay design, we used synthetic ath-miR159a as the template. Synthetic RNA oligo was quantified based on the A260 value and diluted over seven orders of magnitude. As shown in Figure 2a, PCR on the 10-fold dilution series generated concentration-dependent amplification curves whose exponent phases were parallel and about five cycles apart between each 10-fold change. This method allows us to detect as little as 0.0001 fg (seven copies per reaction) of synthetic ath-miR159a. Our results of small RNA purification indicated that on average every 1 ng of total RNA contains about 0.02 pg of ath-miR159a in Arabidopsis leaves. Therefore, our method is capable of measuring small RNAs in biological samples. The CT values for each of the dilutions were plotted against input concentrations of synthetic ath-miR159a oligo, and the result showed excellent linearity between the log of input amount and CT value (Figure 2b). The slope of the standard curve was –3.47, suggesting that the efficiency of the reaction was above 94% with a high correlation coefficient of 0.999.

Figure 2.

 TaqMan quantitative real-time PCR on synthetic ath-miR159a. (a) Amplification plot of synthetic ath-miR159a over seven orders of magnitude. The amount of synthetic ath-miR159a was calculated based on OD260 nm and a 10-fold dilution series was made. Estimated input in the reverse transcription step ranged from 1.3 × 10−3 fm to 1.3 × 104 fm, equivalent to 7 to 7 × 107 copies per reaction, respectively. (b) Standard curve of the synthetic ath-miR159a.

When evaluating our design, we focused on three critical parameters: sensitivity, dynamic range and specificity. Sensitivity was defined as the cycle threshold (CT) value at which an assay containing 1.3 fm of synthetic miRNA was detected. The lower this CT value, the more sensitive was the assay. Dynamic range was defined as the CT difference between 13 fm amount template spike-in (signal) and a no-template control (background). As shown in Figure 3, the average CT value was 22.23 when the assay was conducted following our design; 2.71 cycles faster than an average of 24.94 when the primers and probe were prepared according to the previous design (Chen et al., 2005). The dynamic range of our method was 16.62 (38.85–22.23) while it was 13.37 (38.31–24.94) when the assay was conducted according to the previous design (Chen et al., 2005). Assays with a high dynamic range allow measurements of very low miRNA/siRNA copy numbers. We think that the better sensitivity and greater dynamic range are partially because that our probe was shorter and designed to directly anneal to the cDNA product while that of Chen et al. to the amplicon.

Figure 3.

 Dynamic range and sensitivity of TaqMan PCR-based ath-miR159a assay. TaqMan real-time PCR was conducted using 7 × 103 copies (a and b) of synthetic RNA or no template control (a′ and b′) according to either the present (a and a′) or the previous design (b and b′).

To achieve better specificity, we placed the reporter dye on the nucleotide complementary to the small RNA, instead of the nucleotide complementary to the stem-loop primer (Chen et al., 2005). To evaluate the specificity, we designed and purchased ath-miR159a variants which differ by one to three nucleotides compared to the wild-type ath-miR159a (Figure 4a). Using these RNA oligos as templates and the reverse transcription primer specific to the wild type, we synthesized cDNAs. Then we ran real-time PCR using the cDNA products as templates with the forward primer and the TaqMan probe that were designed based on the wild-type sequence. As shown in Figure 4b, dependent on the location, even one mismatch dramatically reduced the relative detection down to zero, suggesting that this method can discriminate among related small RNAs that differ by as little as one nucleotide. More interestingly, the relative detection was only 20% when the RNA duplex was used as the template, suggesting that this method will not be affected by either primary miRNA or precursor miRNA and, therefore, is suitable for quantification of mature miRNA. This will be useful to understand the biogenesis and functions of miRNA.

Figure 4.

 Specificity of the TaqMan ath-miR159a assay. (a) Sequences of ath-miR159a, its duplex and variants. Mismatches are printed in red. (b) Relative detection of ath-miR159a duplex (DX) and variants (V1~V7) using the primers and probe designed for the wild type ath-miR159a. A total of 7 × 103 copies of synthetic ath-miR159a, its duplex or variants were used for each reverse transcription reaction. Relative detection (%) calculated based on the CT difference between perfectly matched and mismatched targets as relative detection (%) = 2−ΔCT × 100%; where ΔCT = CT variant – CT wild type.

Expression analysis of miRNA using the TaqMan-based real-time PCR method

In an effort to identify novel miRNAs, we are cloning and sequencing small RNAs prepared from plants using the 454 technology (Margulies et al., 2005). gma-miRMON1 is new miRNA isolated from soybean leaves. Using the quantitative real-time PCR method described above, we investigated the expression pattern of gma-miRMON1. As shown in Figure 5a, this miRNA was expressed 10-fold higher in the leaves than in the roots. The quantitative data are consistent with the Northern blotting result (Figure 5b), indicating that our quantitative real-time PCR method can accurately measure mature miRNA levels in total RNA isolated from biological samples. The high-throughput nature of the real-time PCR method encouraged us to apply it to transcriptional profiling of the miRNA in various tissues at different stages. As shown in Figure 6a, gma-miRMON1 was profoundly expressed in the leaves, especially in the mature leaves, which is well correlated to the precursor miRNA levels determined by the Affymetrix microarray (Figure 6b), suggesting that the quantitative real-time PCR method can be a powerful tool to validate transcriptional profiling results.

Figure 5.

 Expression of gma-miRMON1. (a) TaqMan quantitative real-time PCR data. Relative expression levels of gma-miRMON1 were determined using the present TaqMan real-time PCR method. Each assay contained 50 ng of total RNA isolated from leaves and roots with the Ambion kit. Error bars correspond to one standard deviation of three duplicates. (b) Northern blotting result. Each lane was loaded 5 µg of the total RNA and probed with the digoxigenin-labelled DNA oligo complementary to gma-miRMON1.

Figure 6.

 Expression profiles of gma-miRMON1. (a) Mature gma-miRMON1 levels across a panel of 39 tissues were determined by the TaqMan quantitative real-time PCR. Each assay contained 50 ng of total RNA isolated by the Ambion kit. (b) gma-miRMON1 precursor levels in the same panel of RNA samples. Total RNA (5 µg) was applied to Affymetrix microarray using a custom soybean GeneChip. Error bars correspond to one standard deviation of three duplicates.

Quantification of siRNA using the TaqMan-based real-time PCR method

To test if the quantitative real-time PCR method can be used to quantify siRNA in transgenic lines, we extracted total RNA from two transgenic lines carrying a RNAi transgene against corn rootworm. We predicted potential siRNAs from the sequence of the RNAi repeat using an online program SIRNA from EMBOSS (http://bioweb.pasteur.fr/seqanal/interfaces/sirna.html#outseq). We then designed an assay for one siRNA and ran the quantitative real-time PCR. As shown in Figure 7b, the siRNA level was 2-fold higher in Line 1 than in Line 2, which was correlated to the transgene transcript levels determined by the conventional TaqMan real-time PCR on the 3′UTR derived from the RNAi transgene (Figure 7a) and perfectly matched the Northern blotting result (Figure 7c).

Figure 7.

 Expression of a vATPase siRNA. (a) Transgene expression levels. Transcripts derived from the transgene were quantified by the conventional TaqMan real-time PCR using the primers and probe specific to the 3′UTR of the heat shock protein 17 gene (HSP17) that was attached to the reverse repeat against corn rootworm vATPase. Each assay contained 50 ng of total RNA that was isolated from leaves of two independent lines with the Ambion kit. Error bars correspond to one standard deviation of three duplicates. (b) Quantification of the vATPase siRNA (CCAAGCCAGUCUACUGAAGGG). Relative expression levels of a vATPase siRNA were determined using the present TaqMan real-time PCR method. Each assay contained 50 ng of the total RNA. Error bars correspond to one standard deviation of three duplicates. (c) Northern blotting result. Each lane was loaded 5 µg of the total RNA and probed with the digoxigenin-labelled DNA oligo complementary to the siRNA.


siRNAs, miRNAs, and other evolutionarily conserved non-coding RNA are involved in post-transcriptional gene regulation. Abnormal non-coding RNA levels can cause serious healthy problems. For example, deregulated miRNA levels are observed in human diseases, including cancer and neurodegeneration. To further understand biogenesis and functions of these small RNAs and to develop new siRNA expression systems for a variety of genetic and therapeutic applications, optimized purification and detection methods are required. A SYBR Green-based method was first developed for quantifying miRNA precursors (Schmittgen et al., 2004). However, this method cannot detect mature miRNAs. Therefore, different efforts have been made to develop real-time PCR for small RNA quantification (Table 1).

A poly(T) adapter was used to convert miRNAs to cDNAs in an SYBR Green-based method for quantifying miRNAs (Shi and Chiang, 2005). miRNAs were polyadenylated and reverse-transcribed with the poly(T) adapter into cDNAs for real-time PCR using the miRNA-specific forward primer and the sequence complementary to the poly(T) adapter as the reverse primer. This method, using as little as 100 pg total RNA, could readily discriminate the expression of miRNAs having as few as one nucleotide sequence difference. The advantage of this method is that the products of one reverse transcription reaction represent all miRNAs in the RNA sample. The disadvantages include biases derived from the efficiency of polyadenylation, especially for plant miRNAs which are methylated, and the competition for the adapter between the polyadenylated miRNAs and regular mRNAs which are overwhelmingly more in total RNA. Moreover, phenol–chloroform extraction and ethanol precipitation steps do not allow high-throughput operation.

To overcome these disadvantages, a simple and robust assay for quantitative analysis of miRNAs was developed. The method again relies on reverse transcription followed by SYBR Green-based real-time PCR, but it uses an extension primer containing miRNA-specific sequence for reverse transcription and an LNA-modified miRNA-specific reverse primer for PCR (Raymond et al., 2006). LNA possess a 2′-O, 4′-C methylene bridge in the ribose moiety of nucleotide (Petersen and Wengel 2003). The modification stabilizes the conformation of the sugar group and thereby increases the melting temperatures of oligonucleotides that contain LNA bases. Therefore, this method is capable of discriminating between related miRNA family members that differ by subtle sequence differences.

Since SYBR Green I dye chemistry will detect all double-stranded DNA, including contaminated genomic DNA and non-specific reaction products, a TaqMan-based real-time PCR method was developed (Chen et al., 2005). This method uses stem-loop primers for reverse transcription. The method is specific for mature miRNAs and can discriminate between related miRNAs that differ by as little as one nucleotide. The method needs as little as 25 pg of total RNA for most miRNAs from less than 10 to more than 30 000 copies per cell and exhibits a dynamic range of seven orders of magnitude.

Stem-loop reverse transcription primers are better than conventional ones in terms of reverse transcription efficiency and specificity. In the present article, we described the key-like primers that are even better than the stem-loop primers for small RNA-specific reverse transcription and a new robust TaqMan-based real-time PCR method. By using the synthetic miRNAs, the plant endogenous miRNA, and the siRNA from the transgenic plants, we demonstrated this method could quickly and accurately quantify both miRNAs and siRNAs. Compared with existed real-time PCR assays, the new method showed higher sensitivity and better dynamic range with most assays allowing measurements in the femtomolar range. Therefore, it enables miRNA expression profiling to identify and monitor potential biomarkers specific to tissues or diseases. In conclusion, the method provides a valuable tool for small RNA researches in all organisms, including plants and animals.

Experimental procedures

RNA preparation

Total RNA including small RNA was extracted using the mirVana™ miRNA Isolation Kit (Ambion, Austin, TX, USA) following the recommended protocol. The kit, which uses organic extraction followed by purification on a glass fibre filter using specialized binding and wash solutions, effectively recovers all RNA – from large mRNA and ribosomal RNA down to 10 mers – from virtually all cell and tissue types. For assessment of sensitivity and specificity of the real-time PCR, RNA oligos and duplex were synthesized and used as templates.

Primers and TaqMan probes

For small RNA-specific cDNA synthesis, a reverse transcription primer was designed for each small RNA tested. For small RNA-specific amplification, a forward primer identical to the 5′ end of a small RNA, 14–18 nucleotides in length, was designed. The reverse primer is 19 nucleotides long (5′-CGCGAGCACAGAATTAATA-3′). It is universal for all small RNAs and complementary to the 3′ end of the reverse transcription primer. and a TaqMan probe was designed for each small RNA as FAM 5′-NNNCGCGTATAGTGAGT-3′ TAMRA. The 5′-NNN is complementary to part of the small RNA and the CGCGTATAGTGAGT-3′ to part of the reverse transcription primer. For an endogenous reference, 18S was selected. The forward and reverse primers and the TaqMan probe were 5′-CGTCCCTGCCCTTTGTACAC-3′, 5′-CGAACACTTCACCGGATCATT-3′ and VIC 5′-CCGCCCGTCGCTCCTACCGAT-3′ TAMRA, respectively. All DNA and RNA oligos were ordered from the Integrated DNA Technologies (Coralville, IA, USA). All probes were purchased from Applied Biosystems (Foster City, CA, USA). See Table 2 for a complete list of primers and probes used in this study.

Table 2.   Primers and probes used in this study
vATPase siRNA01
ABI design for ath-miR159a

cDNA synthesis

For cDNA synthesis, up to 25 pg/5 µL of RNA sample, 0.05 µL of 100 µm reverse transcription primer, 0.01 µL of 100 pm 18S reverse primer and 0.25 µL of 10 mm dNTP mixture, and 1.69 µL of distilled water were mixed using a 96-well plate and denatured at 65 °C for 5 min in a Thermal Controller (MJ Research, Watertown, MA, USA). The mixture was kept at 4 °C for 5 min before adding 2 µL of 5× First-strand Buffer, 0.5 µL of 0.1 m DTT, 0.25 µL of RNaseOUT™ Recombinant Ribonuclease Inhibitor (40 units/µL) and 0.25 µL of SuperScript® III Reverse Transcriptase (200 units/µL). All these reagents were purchased from Invitrogen (Carlsbad, CA, USA). The solution was mixed well, incubated at 16 °C for 30 min and at 42 °C for 30 min to synthesize cDNA, and denatured at 70 °C for 15 min to inactivate the enzymes in a Thermal Controller from MJ Research. cDNA products were diluted 100 times and stored at –20 °C if not used right away.

Real-time PCR and data analysis

Diluted cDNA products (5 µL) were added to the mixture of 0.015 µL each of 100 µm 18S forward primer, 18S reverse primer and 18S probe, 0.03 µL each of 100 µm small RNA-specific forward primer and the TaqMan probe, the universal reverse primer, and 4.865 µL 2×TaqMan Master mix (Applied Biosystems) using a 384-well plate. Real-time PCR was conducted in 10 µL of total volume on ABI 7900HT at 95 °C for 10 min followed by 40 cycles of 95 °C for 15 sec and 48 °C for 1 min. The threshold cycles (CT) for 18S and small RNA were determined by ABI program SDS 2.1 and exported to an Excel file for further calculation. The relative expression level was defined as Power (2, –ΔΔCT), where ΔΔCT = CT small RNA– CT 18S–[40-Average CT 18S]. Average CT of 18S was calculated from all samples.

Northern blotting on small RNA

Five microgram of total RNA from each sample was denatured at 95 °C for 4 min, loaded on a 17% polyacrylamide gel containing 7 m urea and electrophoresed at 180 volts in 0.5× TBE buffer until the BPB dye reaches the bottom of the gel. The separated RNA was transferred onto a nylon membrane (Schleider and Schuell, Dassel, Germany) in a Trans-blot SD Semi-dry Transfer Cell (Bio-Rad, Hercules, CA, USA) at 400 mAmps for 1 h. The membrane was air-dried and auto-crosslinked at 1200 µJ × 100 in a Stratalinker 1800 (Stratagene, Cedar Creek, TX, USA) and probed using the digoxigenin-labelled DNA oligo complementary to small RNAs in the PerfectHyb Plus buffer (Sigma, St. Louis, MO, USA) overnight at 38°C in a hybridization oven. Post-hybridization washes and hybridizing signal visualization were conducted following the Instructional Manual for the DIG Northern Starter Kit (Roche, Indianapolis, IN, USA).


Transcriptional profiling was conducted using the Affymetrix microarray platform (Santa Clara, CA, USA). Total RNA (5 µg) from each samples was used to synthesize cDNA. Biotin-labelled cRNA was prepared from the cDNA samples, fragmented and then loaded to the custom soybean GeneChips, which were ordered from Affymetrix. Each gene had six probes covering different regions on the high-density chip. Hybridization, post-hybridization washing, image scanning and data analysis were conducted following the Expression Analysis Technical Manual. The difference in perfect and mismatch probe intensities after normalization is used for gene expression measurements.


We thank Matt M. Tanzer, Jintai Huang, and Dafeng Zhou for providing plant materials; Mary M. Blanchard, John E. McLean, Tom H. Adam and Steve Padgette for their leadership; and our team members for useful discussions.