Downloaded from rnajournal.cshlp.

org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

BIOINFORMATICS

miPrimer: an empirical-based qPCR primer design method

for small noncoding microRNA

SHIH-TING KANG,1 YI-SHAN HSIEH, CHI-TING FENG, YU-TING CHEN, POK ERIC YANG, and WEI-MING CHEN1
Quark Biosciences, Zhubei, Hsinchu, 30261, Taiwan

ABSTRACT
MicroRNAs (miRNAs) are 18–25 nucleotides (nt) of highly conserved, noncoding RNAs involved in gene regulation. Because
of miRNAs’ short length, the design of miRNA primers for PCR amplification remains a significant challenge. Adding to the
challenge are miRNAs similar in sequence and miRNA family members that often only differ in sequences by 1 nt. Here,
we describe a novel empirical-based method, miPrimer, which greatly reduces primer dimerization and increases primer
specificity by factoring various intrinsic primer properties and employing four primer design strategies. The resulting primer
pairs displayed an acceptable qPCR efficiency of between 90% and 110%. When tested on miRNA families, miPrimer-
designed primers are capable of discriminating among members of miRNA families, as validated by qPCR assays using Quark
Biosciences’ platform. Of the 120 miRNA primer pairs tested, 95.6% and 93.3% were successful in amplifying specifically
non-family and family miRNA members, respectively, after only one design trial. In summary, miPrimer provides a cost-
effective and valuable tool for designing miRNA primers.
Keywords: microRNA; primer design strategies; primer dimerization; primer specificity; real-time PCR; qPCR efficiency

INTRODUCTION ods, such as small RNA sequencing, northern blot analysis,

in situ hybridization, nanoparticle-based methods, miRNA
miRNAs, short endogenous noncoding RNAs consisting of
qPCR assays, and miRNA microarrays, are routinely used
18–25 nucleotides (nt), are involved in post-transcriptional
in the detection and expression analysis of miRNAs
regulation by targeting the 3′ UTR of mRNA for mRNA deg-
(Catuogno et al. 2011; Tian et al. 2015). Among these meth-
radation or translation inhibition (Lee et al. 1993; Wightman
ods, qPCR assay is one of the most widely used measurement
et al. 1993; Reinhart et al. 2000; Bartel 2009). First identified in
methods for miRNA quantification and expression profiling
Caenorhabditis elegans, they function to regulate the develop-
because of its sensitivity, convenience, and short experimen-
ment timing in larval stage (Lee et al. 1993; Wightman et al.
tal time. Several primer and probe design methods have been
1993). Subsequently, over 35,828 mature miRNAs were found
developed for qPCR-based miRNA quantification, such as
in 223 species, including plants, animals, and viruses (http://
the stem–loop RT-qPCR method with a hydrolysis probe
www.mirbase.org/).
(Chen et al. 2005; Feng et al. 2009), SYBR Green RT-qPCR
miRNAs are key regulators of gene expression networks,
method with LNA primers (Raymond et al. 2005), and
controlling diverse biological processes including prolifera-
poly(A)-tailed universal reverse transcription-based method
tion, differentiation, apoptosis, metabolism, development,
with a specific forward primer and a universal reverse primer
and host−pathogen interactions (Bueno et al. 2008; Slaby
or two specific forward and reverse primers (Fu et al. 2006;
et al. 2009; Small and Olson 2011). Furthermore, miRNAs
Balcells et al. 2011). Because mature miRNA are short, some-
have been demonstrated to be involved in cancer develop-
times between 18 and 25 nt long, one of the obstacles in a
ment (Calin and Croce 2006; Esquela-Kerscher and Slack
miRNA qPCR assay is to design highly specific primers
2006; Manikandan et al. 2008; Zhang et al. 2014; Ohtsuka
with a Tm of 59°C–60°C and acceptable PCR efficiency.
et al. 2015) and the progression of other diseases (Lu et al.
While hydrolysis probes and LNA primers both met the
2008; Jiang et al. 2009; Li and Kowdley 2012; Mendell and
criteria, they are expensive. Balcells et al. (2011) provides a
Olson 2012).
cost-effective method using two specific primers, with in-
The implications of miRNAs in cancer and other diseases
creased Tm and specificity by adding a 5′ -tail to each primer.
have led to significant efforts and resources divested in the
In addition, the authors have shown that the method has a
understanding of miRNA expression. To date, many meth-
higher amplification efficiency than LNA primers, and could
1
These authors contributed equally to this work.
Corresponding author: [email protected] © 2018 Kang et al. This article, published in RNA, is available under a
Article is online at http://www.rnajournal.org/cgi/doi/10.1261/rna.061150. Creative Commons License (Attribution 4.0 International), as described at
117. Freely available online through the RNA Open Access option. http://creativecommons.org/licenses/by/4.0/.

304 RNA 24:304–312; Published by Cold Spring Harbor Laboratory Press for the RNA Society
 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

miPrimer: qPCR primer design method for miRNA

successfully discriminate closely related miRNAs as well as methodology (Fig. 1) can achieve excellent qPCR efficiency,
LNA primers. However, current poly(A)-tailed universal a titration assay of four 10-fold serial dilutions was conducted
reverse transcription-based methods do not provide a sys- with different miRNAs in various designing methods
tematic way of designing primers to distinguish between (Materials and Methods). First, a miRNA serial titration assay
members of miRNA families. was performed using synthetic oligonucleotide hsa-miR-9-
In the present study, we established an algorithm called 5p as a template, a forward primer hsa-miR-9-5p-F, and a
miPrimer, which significantly improves the way primers are universal reverse primer (pair of primers designed by the
designed in the poly(A)-tailed universal reverse transcription- “uni-system,” see Fig. 1). As shown in Figure 2A, hsa-mir-
based method. Using Quark Biosciences’ platform (Chang Y, 9-5p demonstrated a qPCR efficiency of between 90% and
Wei CW, Pan CC, Chiou CF, Chang CH, Hsieh YF, Lee CH, 110% (N = 3, STD ≤ 0.41), which is considered acceptable
Huang JW, Zheng XY, Chiu CY, et al., in prep.), we have dem- (Robledo et al. 2014). To evaluate the qPCR efficiency of
onstrated that the design method produces primers that are ca- the “specific-FR-system,” oligonucleotide hsa-let-7b-5p, a
pable of discriminating closely related miRNAs as well as member of the let-7 miRNA family, was selected as a tem-
miRNA family members with multiple- or single-nucleotide plate for the miRNA serial titration assay. The primer pairs,
difference(s). To our knowledge, miPrimer is the first and hsa-let-7b-5p-F and hsa-let-7b-5p-R, were designed by
only empirical primer design method that provides research “specific-FR-systemoverlap” (Fig. 1). The result illustrated an
scientists a cost-effective, systematic way of designing miRNA acceptable qPCR efficiency of between 90% and 110% (Fig.
primers with a high success rate. 2B, N = 3, STD ≤ 0.1). To demonstrate that miPrimer can
design primer pairs with acceptable qPCR efficiency for
templates other than miRNA, we constructed an artificial
RESULTS
template to be tested. The result of the serial titration assay
indicated that the primer pairs for the artificial template
Achieving excellent qPCR efficiency with miPrimer
had a qPCR efficiency of 94.95% (Fig. 2, N = 3, STD ≤
The qPCR efficiency is impacted by a number of factors. To 0.07). We also performed titration assays for two other prim-
illustrate that the miRNA primers designed by miPrimer er sets, has-miR-122-5p (uni-system) and has-miR-10a-5p

FIGURE 1. Framework of miPrimer’s design strategies. miPrimer includes two major systems: uni-system and specific-FR-system. The specific-FR-
system contains three sequential design rules/functions, overlap, FPM (forward primer major), and RPM (reverse primer major), to reduce dimer
issues and increase primer specificity. There are four design strategies highlighted in color, from left to right, uni-system, specific-FR-systemoverlap,
specific-FR-systemFPM, and specific-FR-systemRPM. A more detailed description of when and how to utilize the strategies is transcribed in the
Materials and Methods section.

www.rnajournal.org 305
 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

Kang et al.

evaluation. A qPCR reaction was per-

formed using synthetic hsa-miR-18
oligonucleotide, which mimics as tem-
plates, and primer pairs designed by the
specific-FR-systemoverlap (Fig. 3A). The
result indicated that the synthetic mimics
can be successfully discriminated by their
respective primer pairs, either the for-
ward primer hsa-miR-18a-5p-F/reverse
primer hsa-miR-18a-5p-R set or the for-
ward primer hsa-miR-18b-5p-F/reverse
primer hsa-miR-18b-5p-R set (Fig. 3A).
Despite the range of the Cq value (this
type of range is often observed in nano-
volume qPCR platforms as described in
Farr et al. 2015), the experiment served
the purpose of illustrating the specificity
of the primer pairs. In another illustra-
FIGURE 2. qPCR efficiency of miRNA primers designed by miPrimer. (A) qPCR efficiency anal-
ysis of a serially diluted oligonucleotide hsa-miR-9-5p miRNA template against hsa-miR-9-5p-F tion, the primers FRM
designed by specific-
primer and universal reverse primer designed by uni-system (four 10-fold dilutions; 2500 qPCR FR-system for hsa-miR-16 family
reactions per dilution). (B) qPCR efficiency analysis of a serially diluted synthetic oligonucleotide members, hsa-miR-16-5p and hsa-miR-
hsa-let-7b against hsa-let-7b-F/R designed by specific-FR-systemFPM (four 10-fold dilution; 2500 195-5p, can distinguish synthetic hsa-
qPCR reactions per dilution). (C) qPCR efficiency analysis of a serially diluted artificial template
using the universal reverse primer and a designed forward primer (four 10-fold dilutions; 2500 miR-16 family oligonucleotide template
qPCR reactions per dilution). Each data point on the plot represents mean Cq ± SD from three mimics (Fig. 3B).
replicates. The hsa-let-7 family, consisting of
eight miRNAs where four of the eight
have almost identical sequences, creates
(“specific-FR-system”). The results showed that both as- one of the biggest challenges in primer designs. Our design
says had acceptable qPCR efficiencies (Supplemental Fig. process was able to produce primers that distinguish hsa-
S1). Moreover, the same five pairs of primers performed let-7 family members, achieving up to <5% cross-reactivity
on a Bio-Rad qPCR platform showed similar results between one or more nucleotide mismatches (Fig. 3C).
(Supplemental Fig. S2), albeit with a slightly lower PCR effi- Only the hsa-let-7c miRNA assay showed 8% cross-reactivity
ciency. We believe the reason for a slight decrease in the PCR against a synthetic hsa-let-7a-5p miRNA template (Fig. 3C).
efficiency of the Bio-Rad system is due to the qPCR Master Finally, we excluded nonspecific binding by performing no-
Mix, which was specific and optimized only for Quark template control (NTC) experimental assays. One exception
Biosciences’ platform. Taken together, the results suggest was hsa-miR-18a, which has a false positive rate close to 2%
that the miRNA primer pairs designed by either the uni-sys- (Fig. 3D). False positive rate is calculated by PNTC/N, where
tem or specific-FR-system have acceptable qPCR efficiency PNTC denotes the number of reactions in NTC assay, and N
and can be used on various qPCR platforms. denotes the total number of qPCR reaction assays (in our
case, 2500 reactions). In addition, to illustrate that the prim-
ers will distinguish cDNA product from miRNA synthetic
RNA template, we performed reverse transcription on
Discriminative identification of miRNA family
the template followed by qPCR analysis. As shown in
using miPrimer
Supplemental Table S2, the hsa-let-7b primer pair recog-
Discriminating between members of same miRNA family nized only the cDNA synthesized from hsa-let-7b synthetic
with closely related sequences in the qPCR assay represents RNA template, but not other synthesized cDNA. In conclu-
a significant challenge. Our empirical-based method sion, the results demonstrated that miPrimer primers are ca-
miPrimer (Fig. 1) is capable of distinguishing members pable of discriminating members of the miRNA family in
of the same miRNA family by increasing primer specificity qPCR reactions through the increase of primer specificity
while reducing the primer dimer issue. To evaluate the and the elimination of the primer dimer issue.
capability of the primers designed by miPrimer in discrimi- The specificity of the primers was also tested on A2058 cell
nating the miRNA family, a number of assays were designed line miRNAs using the Bio-Rad qPCR platform (Supplemen-
for evaluation. The hsa-miR-18 family, consisting of two tal Table S3). Based on the melting curve analysis, we ob-
miRNAs hsa-miR-18a-5 and hsa-miR-18b-5p that differed served only a single peak for each primer set tested. This
in sequences by a single nucleotide, was used in an illustrates that the primer pairs work not only on synthetic

306 RNA, Vol. 24, No. 3

 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

miPrimer: qPCR primer design method for miRNA

FIGURE 3. Discrimination of miRNA family with primers designed by miPrimer. (A) Mature miRNA sequence of hsa-miR-18 family members.
Sequence differences between members are indicated in red. Primers were designed by specific-FR-systemoverlap. Discrimination of synthetic oligo-
nucleotide hsa-miR-18 family members against homologous or heterologous miRNA primers was performed in 2500 qPCR reaction assays with
Quark Biosciences’ DigiChip. Cq distributions of synthetic oligonucleotide hsa-miR-18 family members against homologous or heterologous
miRNA were plotted based on the result of qPCR reactions. (B) Mature miRNA sequence of hsa-miR-16 family members. Sequence differences be-
tween members are indicated in red. Primers were designed by specific-FR-systemFPM. Discrimination of synthetic oligonucleotide hsa-miR-16 family
members against homologous or heterologous miRNA primers was performed in 2500 qPCR reaction assay with Quark Biosciences’ DigiChip. Cq
distributions of synthetic oligonucleotide hsa-miR-16 family members against homologous or heterologous miRNA were plotted based on the result
of qPCR reactions. (C) Mature miRNA sequence of hsa-let-7 family members. Sequence differences between members are indicated in red. (D) No-
template control (NTC) assay for each miRNA primer was performed to estimate the false positive rate.

templates but also on miRNA templates from real samples. Table 1 shows the efficiency of designing primer based on
Further, we added additional evidence that the primer design the principles of miPrimer. In our case, the less the number
could work across different qPCR platforms. of times a primer pair needs to be redesigned, the more effi-
cient the process is. As shown in the Supplemental Table S2,
95.6% of the primer pairs for 90 non-family miRNAs were
Efficiency of designing miRNA primer via miPrimer
successfully designed by miPrimer with the uni-system in
Designing primers for miRNAs is time-consuming and cost- the first trial. The remaining 4.4% of the primer pairs were
ly, especially for members of miRNA family or miRNAs with redesigned utilizing the specific-FR-system to address the
similar sequences. A total of 120 miRNA primer pairs were dimer issue (Table 1). In addition to the 90 non-family
designed by miPrimer (Table 1) and validated with synthetic miRNAs, there were 17 miRNA families (or similar miRNA
oligonucleotide template mimics on Quark Biosciences’ groups), consisting of 30 total miRNAs, whose primer
qPCR platform (data not shown). A statistical analysis in pairs have been designed by miPrimer. Of the 30 miRNAs,

www.rnajournal.org 307
 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

Kang et al.

TABLE 1. Performance of miPrimer method

Success at first design trial Success at second design trial

Non-family Family Non-family Family

hsa-miR-2114-3p hsa-miR-326 hsa-let-7a-5p hsa-miR-486-5p hsa-miR-30a-5p

hsa-miR-222-3p hsa-miR-335-5p hsa-let-7b-5p hsa-miR-210-3p hsa-miR-520b
hsa-miR-143-3p hsa-miR-338-5p hsa-let-7c-5p hsa-miR-411-5p
hsa-miR-17-3p hsa-miR-361-5p hsa-miR-103a-3p hsa-miR-589-5p
hsa-miR-202-3p hsa-miR-372-3p hsa-miR-16-5p
hsa-let-7d-3p hsa-miR-451a hsa-miR-195-5p
hsa-miR-101-3p hsa-miR-423-5p hsa-miR-23a-3p
hsa-miR-122-5p hsa-miR-378a-5p hsa-miR-30d-5p
hsa-miR-1254 hsa-miR-382-5p hsa-miR-30e-5p
hsa-miR-125b-5p hsa-miR-409-3p hsa-miR-100-5p
hsa-miR-126-3p hsa-miR-425-3p hsa-miR-10a-5p
hsa-miR-1290 hsa-miR-452-3p hsa-miR-10b-5p
hsa-miR-142-3p hsa-miR-483-5p hsa-miR-133a-3p
hsa-miR-145-5p hsa-miR-484 hsa-miR-133b-3p
hsa-miR-150-5p hsa-miR-499a-5p hsa-miR-135b-5p
hsa-miR-151a-3p hsa-miR-425-5p hsa-miR-135a-5p
hsa-miR-152-3p hsa-miR-574-3p hsa-miR-141-3p
hsa-miR-155-5p hsa-miR-574-5p hsa-miR-200a-3p
hsa-miR-15a-5p hsa-miR-579-3p hsa-miR-18a-5p
hsa-miR-15b-5p hsa-miR-593-5p hsa-miR-18b-5p
hsa-miR-182-5p hsa-miR-596 hsa-miR-215-5p
hsa-miR-183-5p hsa-miR-601 hsa-miR-192-5p
hsa-miR-193a-3p hsa-miR-34a-5p hsa-miR-196a-5p
hsa-miR-1972 hsa-miR-373-3p hsa-miR-196b-5p
hsa-miR-197-3p hsa-miR-375 hsa-miR-376c-3p
hsa-miR-203a-3p hsa-miR-500a-5p hsa-miR-302d-3p
hsa-miR-205-5p hsa-miR-1228-5p hsa-miR-30b-5p
hsa-miR-206 hsa-miR-134-5p hsa-miR-30c-5p
hsa-miR-214-3p hsa-miR-221-3p
hsa-miR-215-5p hsa-miR-223-3p
hsa-miR-21-5p hsa-miR-1-3p
hsa-miR-191-5p hsa-miR-625-5p
hsa-miR-22-3p hsa-miR-652-3p
hsa-miR-224-5p hsa-miR-660-5p
hsa-miR-24-3p hsa-miR-663a
hsa-miR-26a-5p hsa-miR-718
hsa-miR-140-5p hsa-miR-7-5p
hsa-miR-28-3p hsa-miR-760
hsa-miR-299-5p hsa-miR-885-5p
hsa-miR-29a-5p hsa-miR-95-3p
hsa-miR-423-3p hsa-miR-127-3p
hsa-miR-198 hsa-miR-9-5p
hsa-miR-31-5p hsa-miR-128-3p

The table indicates the number of primer pairs that were successfully designed during the first and second trials. The results demonstrate that
miPrimer is a cost-effective and efficient method for designing miRNA primers. Primer sequences are not shown except for those indicated in
Table 2. The miRNA names are obtained from miRBase v21.

only the primer pairs of two miRNAs were required to be DISCUSSION

further adjusted to include a 2-nt overlap between the for-
ward and reverse primers (Table 1). In our current study, The process of designing miRNA primer, requiring both
there were no miRNA primer designs that required the dry laboratory tools and wet laboratory validations, can be
specific-FR-systemRPM, which implied that the dimers were time-consuming and arduous. miPrimer, an empirical-based
frequently caused by the reverse primer. In summary, method, was developed by learning from several failed cases
miPrimer provides an efficient way of designing primers during miRNA primer design phases. Through the fine-tun-
for miRNAs. ing of miPrimer, we were able to come up with an effective

308 RNA, Vol. 24, No. 3

 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

miPrimer: qPCR primer design method for miRNA

TABLE 2. List of miRNA primers

miRNA primer Sequence Self-dimer Hetero-dimer

Universal reverse primer 5′ -CAACTCAGGTCGTAGGCAATTCGT-3′ ΔG = −2.66 –

hsa-miR-9-5p-F 5′ -GCACGCTCTTTGGTTATCTAGCTGTATGA-3′ ΔG = −6.00 ΔG = −4.31
hsa-miR-16-5p-F 5′ -GCACGCTAGCAGCACGT-3′ ΔG = −8.29 ΔG = −4.50
hsa-miR-16-5p-R 5′ -TAGGCAATTCGTTTTTTTTTTTTTTTTTTTTCGCC-3′ ΔG = −3.15
hsa-miR-18a-5p-F 5′ -GCTAAGGTGCATCTAGTGCAGA-3′ ΔG = −8.95 ΔG = −4.95
hsa-miR-18a-5p-R 5′ -TCGTAGGCAATTCGTTTTTTTTTTTTTTTTTTTTCTAT-3′ ΔG = −2.52
hsa-miR-18b-5p-F 5′ -GCTAAGGTGCATCTAGTGCAGT-3′ ΔG = −8.95 ΔG = −4.95
hsa-miR-18b-5p-R 5′ -TCGTAGGCAATTCGTTTTTTTTTTTTTTTTTTTTCTAA-3′ ΔG = −2.52
hsa-miR-195-5p-F 5′ -GCACGCTAGCAGCACAG-3′ ΔG = −8.29 ΔG = −6.03
hsa-miR-195-5p-R 5′ -TAGGCAATTCGTTTTTTTTTTTTTTTTTTTTGCCA-3′ ΔG = −7.56
hsa-let-7a-5p-F 5′ -ACGCTGAGGTAGTAGGTTG-3′ ΔG = −2.20 ΔG = −3.92
hsa-let-7a-5p-R 5′ -GCAATTCGTTTTTTTTTTTTTTTAACTAT-3′ ΔG = −2.06
hsa-let-7b-5p-F 5′ - GCTGAGGTAGTAGGTTGTG-3′ ΔG = −1.12 ΔG = −5.40
hsa-let-7b-5p-R 5′ -ATTCGTTTTTTTTTTTTTTTTTTTTAACCAC-3′ ΔG = −1.35
hsa-let-7c-5p-F 5′ -CACGCTGAGGTAGTAGGTTGTA-3′ ΔG = −2.20 ΔG = −5.94
hsa-let-7c-5p-R 5′ -CAATTCGTTTTTTTTTTTTTTTAACCAT-3′ ΔG = −1.35
hsa-let-7d-5p-F 5′ -ACACAGAGGTAGTAGGTTGC-3′ ΔG = −0.83 ΔG = −2.94
hsa-let-7d-5p-R 5′ -CAATTCGTTTTTTTTTTTTTTTTTTTTAACTATG-3′ ΔG = −1.47

ΔG was calculated by RNAcofold with DNA parameters. The sequences are shown in black, blue, and red, which represent miRNA, poly(A),
and universal sequences, respectively. The miRNA sequences were obtained from miRBase v21.

and economical way of designing miRNA primers. In addi- members. Generally, a stretch of 20 dT residues was used
tion to reasonable qPCR efficiencies, the primer dimer issue in the specific reverse primer with an exact match to the
was greatly reduced and the primers exhibited increased template. If any primer pairs designed by miPrimer cannot
specificity. Of the primer pairs designed for 120 miRNAs, successfully distinguish between miRNA family members
95% were successful in the first trial for either non-family with a cross-reactivity of <5%, the length of the dT in the re-
or family members. Overall, miPrimer is an exceptional verse primer will be decreased to 15 dT residues to reduce
tool for the primer design of small noncoding miRNAs. cross-reactivity. For example, let-7a-5p miRNA assay exhib-
In the process of primer design, ΔG is one of most critical ited high cross-reactivity to the synthetic let-7b-5p using the
factors used to determine the presence of dimers (Shen et al. specific reverse primer with a stretch of 20 dT residues (data
2010). Primers or primer pairs calculated by different not shown). After replacing the reverse primer with 15 dT
algorithms showed different ΔG values. Calculated using residues, the cross-reactivity of the let-7a-5p miRNA assay
OligoAnalyzer 3.1, hsa-miR-16-5p-F and hsa-miR-195-5p- to the synthetic let-7b-5p was reduced to 2% (Fig. 3C).
had a low ΔG value of −10.44 and −10.44 kcal/mol, respec- Another similar example is let-7c-5p. After replacing the
tively. However, the ΔG values of these primers were consid- reverse primer with 15 dT residues, the cross-reactivity of
ered acceptable in miPrimer (Table 2), as validated by the let-7c-5p miRNA primers to synthetic templates of let-
our no-template control experimental assays (Fig. 3D). In 7a-5p and let-7b-5p was reduced to <0.4% (data not shown).
comparison, hsa-miR-18a-5p-F and hsa-let-7b-5p-F, with Therefore, shortening the length of 20 dT residues to at least
extremely low ΔG value as calculated by miPrimer or 15 dT residues in the specific reverse primer might be helpful
OligoAnalyzer 3.1, exhibit false positive signals in the no- in distinguishing between miRNA family members based on
template control reactions, an indication of dimer formation. our experiences.
Therefore, various ΔG determining tools, along with miPrimer, As illustrated from the above discussion, although
should be used when designing primers in our case. miPrimer is a powerful tool, there is still room for improve-
Mestdagh et al. (2014) reported that most miRNA detec- ment in the primer design of miRNA families. Various ΔG
tion methods showed cross-reactivity between let-7 family determining tools and the cross-reactivity ratio are a number
members (hsa-let-7a-5p, hsa-let-7b-5p, hsa-let-7c-5p, and of features that can be used to resolve potential issues. The
hsa-let-7d-5p), with the exception of miRCury (Exiqon’s primer specificity is still a key factor in separating members
LNA primers). As an example, high cross-reactivity was of miRNA families. Identifying additional features for primer
observed in the hydrolysis measurement method in one of design of miRNA families in order to improve the primer
the let-7 family assays. Utilizing the design process of specificity will continue to be a very important task and
miPrimer and Quark Biosciences’ platform (Chang Y, Wei can be incorporated into the miPrimer algorithm. Further,
CW, Pan CC, Chiou CF, Chang CH, Hsieh YF, Lee CH, the methodology cannot distinguish isomiRs. Thus, future
Huang JW, Zheng XY, Chiu CY, et al., in prep.), little or improvements can also be done to the methodology to help
no cross-reactivity was shown between hsa-let-7 family resolve the issues of isomiRs distinction.

www.rnajournal.org 309
 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

Kang et al.

MATERIALS AND METHODS plementary to the 3′ -end of the mature miRNA sequence. By
following the adjustment and confirmation steps, the difference in
Flow schematics of primer design for miRNA the Tm value of both primers should not exceed 2°C, thus eliminat-
ing the issue of primer dimer. If the Tm value of the replacement re-
miPrimer is an empirical-based methodology comprised of two verse primer is <55°C, adding at least one or more nucleotides to the
methods, uni-system and specific-FR-system, for designing primers 5′ -end of the replacement reverse primer from the universal reverse
(Fig. 1). When the sequence similarity of the miRNA of interest is sequence will increase the Tm value to 59°C. Finally, both steps are
<90% of that compared to other miRNAs, the uni-system is first iteratively performed until all conditions are satisfied.
adopted to design primer. The uni-system is preferred as the primers
are easier to design and the process is cost-effective and less time-
consuming. On the other hand, the specific-FR-system is used specific-FR-systemoverlap
when the sequence similarity is higher than 90%, e.g., hsa-let-7 fam-
ily. The specific-FR-system has three embedded functions, overlap, In most situations, primers have difficulties discriminating mem-
FPM (forward primer major), and RPM (reverse primer major), bers of the miRNA families or similar miRNAs differed by a
which aim to reduce primer dimer issue and increase primer speci- single-nucleotide or with sequence similarity >90%. Hence, the spe-
ficity. As a result of these embedded functions, the probability of cific-FR-system is strongly suggested as the preferred process to
discriminating between highly similar miRNAs is increased. In ad- increase the primer specificity and reduce the dimer issue. The spe-
dition to the above-described systems, the methodology requires cific-FR-system contains three sequential design rules/functions: (i)
two sequences depending on the rules and conditions; a 5′ -tailed se- overlap, (ii) FPM, and (iii) RPM to address different problems. To
quence 5′ -CGWTSSRCRC-3′ that will be added to forward primers, begin with, the overlap method is required to perform a sequence
and a universal primer 5′ -CAACTCAGGTCGTAGGCAATTCGT-3′ comparison to determine the location(s) of nucleotide difference
that will act as the reverse primer. Below, we describe when and how between members of miRNA families. The sequence of the forward
the uni-system and the specific-FR-system are applied in different primer generally contains at least the same 4-nt residues of the 5′ -
situations for successful primer design: end of the mature miRNA sequence. The sequence of the reverse
primer generally contains at least 4-nt residues complementary to
the 3′ -end of the mature miRNA sequence. Both the forward and
the reverse primers are designed to overlap by a single nucleotide,
uni-system which is the nucleotide that differs between miRNA family mem-
bers. Similarly, there are two additional steps to ensure the specific-
The uni-system is primarily used to design primers for miRNAs of
ity of the forward and reverse primers. First, an adjustment step is
interest when the miRNA is not a member of a family or when its
performed to ensure that the difference of the Tm value between
sequence similarity is <90% of that compared to other miRNAs.
the forward primer and the reverse primer does not exceed 2°C.
The system utilizes the universal primer and a specific forward
In the case that the Tm value of the forward primer is <55°C, one
primer in most situations. The sequence of forward primer generally
or more nucleotide(s) from the 5′ -tailed sequence is added sequen-
contains the same 12–18 nt of the 5′ -end of the mature miRNA se-
tially to the 5′ -end of the forward primer until a Tm of 59°C is
quence. There are two designing steps in the uni-system to ensure
achieved. Similarly, if the Tm value of the reverse primer is <55°C,
the specificity of forward primer. First, an adjustment step is per-
one or more nucleotide(s) from the universal reverse sequence is
formed to ensure that the difference between the Tm value of the for-
added at the 5′ -end of the reverse primer to increase the Tm value
ward primer and the universal reverse primer does not exceed 2°C.
to 59°C. Secondly, a confirmation step is performed to determine
In the case that the Tm value of the forward primer is <57°C, one or
whether a self-dimer exists and/or hetero-dimer is formed between
more nucleotide(s) from the 5′ -tailed sequence is added sequentially
the forward primer and the reverse primer. As mentioned above, ΔG
to the 5′ -end of the forward primer until a Tm of 59°C is achieved.
is one of the common factors used to check for the presence of
Second, a confirmation step is performed to determine whether
dimers; however, in this study, ΔG of less than −9.0 ± 1 kcal/mol
a self-dimer exists and/or hetero-dimer is formed between the for-
usually causes problems in the qPCR reactions. If ΔG of either
ward primer and the universal reverse primer. Our methodology
the self-dimer or hetero-dimer is greater than −9.0 ± 1 kcal/mol,
uses ΔG, a common characteristic used to assess the presence of
the design of forward and reverse primers is completed. In con-
dimers (Shen et al. 2010); when the ΔG of either the self-dimer or
trast, when ΔG of either the self-dimer or hetero-dimer is less
hetero-dimer is greater than −9.0 ± 1 kcal/mol, no additional mod-
than −9.0 ± 1 kcal/mol, changing the 5′ -tailed sequence to other
ification is needed to be applied to the forward primer. In contrast,
IUPAC alternatives is able to reduce the formation of self-dimer or
when the ΔG of the self-dimer or hetero-dimer is less than −9.0 ± 1
hetero-dimer. Both steps are iteratively performed until all conditions
kcal/mol, either the 5′ -tailed sequence is changed to other IUPAC
are satisfied. If the primer does not exhibit specificity after PCR assays,
alternatives or the dimer-forming nucleotides of the forward primer
the number of overlapping nucleotides can be increased up to 2 nt. If
is removed. Both steps are iteratively performed until the conditions
the primer dimer continues to exist after performing all specific-FR-
are satisfied. In several instances, it has been difficult to eliminate the
systemoverlap modification steps, the specific-FR-systemFPM or specif-
dimer issue by designing and modifying only the forward primer,
ic-FR-systemRPM is then used to eliminate the issue.
e.g., hsa-miR-9-5p. Therefore, it is recommended to modify both the
forward primer and the reverse primer to avoid the dimer region.
The process includes the reduction of the length of the 3′ -end of specific-FR-systemFPM
the forward primer and the replacement of the universal reverse
primer with a redesigned reverse primer. The sequence of the re- If the dimer issue is caused by the reverse primer, the specific-FR-
placement reverse primer generally contains 2–8 nt residues com- systemFPM is suggested to resolve the problem. The concept of the

310 RNA, Vol. 24, No. 3

 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

miPrimer: qPCR primer design method for miRNA

specific-FR-systemFPM is to eliminate the dimer-forming sequences Biosciences’ qPCR platform, where Cqmm denotes Cq of miRNA
of the reverse primer and enhance the specificity of the forward against the heterologous primer and Cqpm denotes Cq of miRNA
primer. The sequence of the forward primer is extended to where against the homologous primer. ΔCq > 4 denotes that the cross-re-
one of the nucleotide differences between the miRNAs is found. activity is <5% calculated by 10ΔCq/S × 100%, where S is the slope of
The sequence of the reverse primer is reduced from its 3′ -end se- the standard curve (Wan et al. 2010).
quences but will need to contain at least 4-nt residues complemen-
tary to the mature miRNA sequences. If the dimer issue still cannot
be resolved, the universal reverse primer is then used to replace the Calculation of false positive rate
reverse primer. Finally, both adjustment and confirmation steps are False positive rate is calculated by PNTC/N, where PNTC denotes the
iteratively performed until all conditions are met. number of reactions in NTC assay, and N denotes the total number
of qPCR reaction assays (in our case, 2500 reactions). In no-tem-
plate control assays, any reaction will be treated as a false positive
specific-FR-systemRPM
even if the Cq value is high.
On the other hand, the RPM method is performed if the dimer issue
is caused by the forward primer. Similarly to the specific-FR-
Synthetic miRNA and cDNA template
systemFPM, the specific-FR-systemRPM is used to resolve the forma-
tion of dimer region in the forward primer and improve the specif- Synthetic single-stranded oligonucleotides, which contain the se-
icity of the reverse primer. The sequence of the reverse primer is quences of the universal RT primer 5′ -CAACTCAGGTCGTAGG
extended to where one of the nucleotide differences is located. CAATTCGTTTTTTTTTTTTTTTTTTTT-3′ , were designed to
The sequence of the forward primer is trimmed from its 3′ -end of mimic the sequences of mature miRNA and miRNA cDNA and
sequence but will contain at least the same 4-nt residues of the ma- used as templates in our study. The oligonucleotides purchased
ture miRNA sequences. Finally, both adjustment and confirmation from Protech Technology Enterprise Co., Ltd. or PURIGO
steps are iteratively performed until all conditions are satisfied. Biotechnology Co., Ltd, were dissolved in TE buffer, aliquoted
The designed primers are currently used on Quark Biosciences’ and stored at −80°C.
miRSCan line of PanelChip products, where multiple miRNAs
can be analyzed on one array chip. The product numbers/names
are PanelChip PCP-01 and PCP-02, and can be found on Quantitative real-time PCR assay using DigiChip and
Quark Biosciences’ miRNA service catalog: http://download. Quark Biosciences’ qPCR platform
quarkbiosciences.com/Brochure/miRNA_list.pdf
For qPCR efficiency assay, five 10-fold serial dilutions were made
from the highest concentration (109 copies/µL) by adding 10 µL
Calculation of ΔG value of the previous solution to 90 µL nuclease-free H2O. Four microli-
ters of each synthetic miRNA serial dilution was added to the qPCR
The ΔG value is calculated by RNAcofold algorithm using the DNA mixtures containing 30 µL of 2× Quarkbio qPCR master mix
parameters (Gruber et al. 2008) and is set at 9.0 ± 1 kcal/mol. (QuarkBiosciences, Inc.), 1.5 µL of 0.25 µM specific forward primer,
OligoAnalyzer 3.1 (Owczarzy et al. 2008) is the most frequently and 1.5 µL of 0.25 µM specific reverse primer or universal reverse
used tool to calculate the ΔG of a primer. Primers with a contro- primer. Twenty-three microliters of nuclease-free water was added
versial ΔG value (e.g., see Table 2, hsa-miR-18b-5p-F with a ΔG to the mixture to a final volume of 60 µL. The master mix was mixed
−8.95 kcal/mol) are additionally confirmed by OligoAnalyzer 3.1 thoroughly and briefly spun down to collect the liquid at the bot-
manually. tom. The master mix was then applied onto DigiChip. All reactions
were run in triplicate.
For qPCR assays demonstrating that miRNA family members can
Calculation of melting temperature be discriminated, 2 µL of the synthetic cDNA template (∼2 × 105
copies) were added to the qPCR mixtures containing 30 µL of 2×
The Tm (melting temperature) value is calculated by the following
Quarkbio qPCR master mix (QuarkBiosciences, Inc.), 1.5 µL of
formula:
0.25 µM specific forward primer, and 1.5 µL of 0.25 µM specific re-
1E3∗H verse primer or universal reverse primer. Twenty-five microliters
Tm = − 2.73.15,
S + 1.987∗ log (C) − log(2E9) of nuclease-free water was added to the mixture to a final volume
of 60 µL. The master mix was mixed thoroughly, briefly spun
where H and S denote the enthalpy and entropy, respectively, and C
down and then applied onto DigiChip.
denotes the initial primer concentration (SantaLucia 1998; von
DigiChip, a 36-mm × 36-mm × 1-mm reaction plate consisting
Ahsen et al. 2001). In miPrimer, the difference of the Tm between
of 2500 wells, was developed to be used with Quark Biosciences’
the forward and the reverse primer should not exceed 2°C ± 0.1.
qPCR platform. To apply the qPCR mixture containing the template
OligoAnalyzer 3.1, is used to confirm the Tm manually as required.
and the primer pairs to be tested, 60 µL of the mixture is dispensed
using a pipetman along the edge of DigiChip. The qPCR mixture
Calculation of cross-reactivity should cover more than the entire length of 50 wells along the
edge. Fifty microliters of the qPCR mixture is then applied across
The number of nucleotides that differ between each other is indicat- the entire surface of the DigiChip via a scraping motion with a glass
ed in the matrix. Cross-reactivity between miRNA family members slide, resulting in 20 nL of master mix per reaction well. The
is shown as the mean of Cq and ΔCq (Cqmm − Cqpm) in Quark DigiChip is then submerged, with reaction wells facing the bottom,

www.rnajournal.org 311
 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

Kang et al.

into a tray containing mineral oil. Each plate is then placed into a Fu HJ, Zhu J, Yang M, Zhang ZY, Tie Y, Jiang H, Sun ZX, Zheng XF.
thermal cycler for the amplification of the template. qPCR was 2006. A novel method to monitor the expression of microRNAs.
performed according to the following cycling program for miRNA Mol Biotechnol 32: 197–204.
Gruber AR, Lorenz R, Bernhart SH, Neuböck R, Hofacker IL. 2008. The
analysis: 95°C for 36 sec to denature for 1 cycle, followed by 95°C
Vienna RNA websuite. Nucleic Acids Res 36: W70–W74.
for 36 sec and 60°C for 72 sec for 40 cycles. Jiang Q, Wang Y, Hao Y, Juan L, Teng M, Zhang X, Li M, Wang G, Liu Y.
Quark Biosciences’ qPCR platform is a real-time quantitative/digi- 2009. miR2Disease: a manually curated database for microRNA
tal PCR Platform developed by Quark Biosciences, Inc. The thermal deregulation in human disease. Nucleic Acids Res 37: D98–D104.
cycling functionality is accomplished by Thermal-Roller-Coaster, a Lee RC, Feinbaum RL, Ambros V. 1993. The C. elegans heterochronic
proprietary technology which consists of six resistive heater blocks gene lin-4 encodes small RNAs with antisense complementarity to
lin-14. Cell 75: 843–854.
with different, but constant temperature. Amplification is achieved
Li Y, Kowdley KV. 2012. MicroRNAs in common human diseases.
by shuttling DNA or cDNA samples between denaturing heater blocks Genomics Proteomics Bioinformatics 10: 246–253.
and annealing/extension heater blocks. Each PCR run can accommo- Lu M, Zhang Q, Deng M, Miao J, Guo Y, Gao W, Cui Q. 2008. An anal-
date up to six samples. The platform utilizes a white-light LED optics ysis of human microRNA and disease associations. PLoS One 3:
system with up to four different filter block formats to achieve illumi- e3420.
nation for samples with FAM, VIC, ROX, and Cy5 dye. Manikandan J, Aarthi JJ, Kumar SD, Pushparaj PN. 2008. Oncomirs: the
potential role of non-coding microRNAs in understanding cancer.
Bioinformation 2: 330–334.
Mendell JT, Olson EN. 2012. MicroRNAs in stress signaling and human
SUPPLEMENTAL MATERIAL disease. Cell 148: 1172–1187.
Mestdagh P, Hartmann N, Baeriswyl L, Andreasen D, Bernard N,
Supplemental material is available for this article. Chen C, Cheo D, D’Andrade P, DeMayo M, Dennis L, et al. 2014.
Evaluation of quantitative miRNA expression platforms in the
microRNA quality control (miRQC) study. Nat Methods 11: 809–815.
ACKNOWLEDGMENTS Ohtsuka M, Ling H, Doki Y, Mori M, Calin GA. 2015. MicroRNA pro-
cessing and human cancer. J Clin Med 4: 1651–1667.
We would like to thank all members of Quark Biosciences Inc. for Owczarzy R, Tataurov AV, Wu Y, Manthey JA, McQuisten KA,
the support of Quark Biosciences’ qPCR platform, which is capable Almabrazi HG, Pedersen KF, Lin Y, Garretson J, McEntaggart NO,
et al. 2008. IDT SciTools: a suite for analysis and design of nucleic
of running the qPCR assay on DigiChip.
acid oligomers. Nucleic Acids Res 36: W163–W169.
Author contributions: W.-M.C. and S.-T.K. designed miPrimer: Raymond CK, Roberts BS, Garrett-Engele P, Lim LP, Johnson JM. 2005.
W.-M.C. wrote the code and S.-T.K. designed and tested primers Simple, quantitative primer-extension PCR assay for direct monitor-
based on the written program. The remaining authors provided ing of microRNAs and short-interfering RNAs. RNA 11: 1737–1744.
guidance and project oversight. W.-M.C. and S.-T.K. wrote the Reinhart BJ, Slack FJ, Basson M, Pasquinelli AE, Bettinger JC,
manuscript. Rougvie AE, Horvitz HR, Ruvkun G. 2000. The 21-nucleotide let-
7 RNA regulates developmental timing in Caenorhabditis elegans.
Nature 403: 901–906.
Received February 16, 2017; accepted November 29, 2017. Robledo D, Hernández-Urcera J, Cal RM, Pardo BG, Sánchez L,
Martínez P, Viñas A. 2014. Analysis of qPCR reference gene stability
determination methods and a practical approach for efficiency cal-
culation on a turbot (Scophthalmus maximus) gonad dataset. BMC
REFERENCES Genomics 15: 648.
SantaLucia J Jr. 1998. A unified view of polymer, dumbbell, and oligo-
Balcells I, Cirera S, Busk PK. 2011. Specific and sensitive quantitative nucleotide DNA nearest-neighbor thermodynamics. Proc Natl Acad
RT-PCR of miRNAs with DNA primers. BMC Biotechnol 11: 70. Sci 95: 1460–1465.
Bartel DP. 2009. MicroRNAs: target recognition and regulatory func- Shen Z, Qu W, Wang W, Lu Y, Wu Y, Li Z, Hang X, Wang X, Zhao D,
tions. Cell 136: 215–233. Zhang C. 2010. MPprimer: a program for reliable multiplex PCR
Bueno MJ, Pérez de Castro I, Malumbres M. 2008. Control of cell pro- primer design. BMC Bioinformatics 11: 143.
liferation pathways by microRNAs. Cell Cycle 7: 3143–3148. Slaby O, Svoboda M, Michalek J, Vyzula R. 2009. MicroRNAs in colo-
Calin GA, Croce CM. 2006. MicroRNA signatures in human cancers. rectal cancer: translation of molecular biology into clinical applica-
Nat Rev Cancer 6: 857–866. tion. Mol Cancer 8: 102.
Catuogno S, Esposito CL, Quintavalle C, Cerchia L, Condorelli G, De Small EM, Olson EN. 2011. Pervasive roles of microRNAs in cardiovas-
Franciscis V. 2011. Recent advance in biosensors for microRNAs cular biology. Nature 469: 336–342.
detection in cancer. Cancers (Basel) 3: 1877–1898. Tian T, Wang J, Zhou X. 2015. A review: microRNA detection methods.
Chen C, Ridzon DA, Broomer AJ, Zhou Z, Lee DH, Nguyen JT, Org Biomol Chem 13: 2226–2238.
Barbisin M, Xu NL, Mahuvakar VR, Andersen MR, et al. 2005. von Ahsen N, Wittwer CT, Schutz E. 2001. Oligonucleotide melting
Real-time quantification of microRNAs by stem-loop RT-PCR. temperatures under PCR conditions: nearest-neighbor corrections
Nucleic Acids Res 33: e179. for Mg2+, deoxynucleotide triphosphate, and dimethyl sulfoxide
Esquela-Kerscher A, Slack FJ. 2006. Oncomirs—microRNAs with a role concentrations with comparison to alternative empirical formulas.
in cancer. Nat Rev Cancer 6: 259–269. Clin Chem 47: 1956–1961.
Farr RJ, Januszewski AS, Joglekar MV, Liang H, McAulley AK, Wan G, Lim QE, Too HP. 2010. High-performance quantification of ma-
Hewitt AW, Thomas HE, Loudovaris T, Kay TW, Jenkins A, et al. ture microRNAs by real-time RT-PCR using deoxyuridine-incorpo-
2015. A comparative analysis of high-throughput platforms for val- rated oligonucleotides and hemi-nested primers. RNA 16: 1436–1445.
idation of a circulating microRNA signature in diabetic retinopathy. Wightman B, Ha I, Ruvkun G. 1993. Posttranscriptional regulation of
Sci Rep 5: 10375. the heterochronic gene lin-14 by lin-4 mediates temporal pattern-
Feng J, Wang K, Liu X, Chen S, Chen J. 2009. The quantification of to- formation in C. elegans. Cell 75: 855–862.
mato microRNAs response to viral infection by stem-loop real-time Zhang W, Liu J, Wang G. 2014. The role of microRNAs in human breast
RT-PCR. Gene 437: 14–21. cancer progression. Tumour Biol 35: 6235–6244.

312 RNA, Vol. 24, No. 3

 Downloaded from rnajournal.cshlp.org on March 27, 2023 - Published by Cold Spring Harbor Laboratory Press

miPrimer: an empirical-based qPCR primer design method for small

noncoding microRNA
Shih-Ting Kang, Yi-Shan Hsieh, Chi-Ting Feng, et al.

RNA 2018 24: 304-312 originally published online December 5, 2017

Access the most recent version at doi:10.1261/rna.061150.117

Supplemental http://rnajournal.cshlp.org/content/suppl/2017/12/05/rna.061150.117.DC1
Material

References This article cites 32 articles, 4 of which can be accessed free at:
http://rnajournal.cshlp.org/content/24/3/304.full.html#ref-list-1

Open Access Freely available online through the RNA Open Access option.

Creative This article, published in RNA, is available under a Creative Commons License
Commons (Attribution 4.0 International), as described at
License http://creativecommons.org/licenses/by/4.0/.

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the
Service top right corner of the article or click here.

To subscribe to RNA go to:

http://rnajournal.cshlp.org/subscriptions

© 2018 Kang et al.; Published by Cold Spring Harbor Laboratory Press for the RNA Society