Manipulating MicroRNAs

Table of Contents

Letter from the Editor
Index of Experts
Q1: How do you effectively isolate and purify miRNA from cells?
Q2: How do you efficiently clone an miRNA?
Q3: What method do you use to ensure sensitive and specific gene expression profiling of miRNAs? Why?
Q4: How do you validate your expression results so that you get robust and reproducible data?
Q5: What method do you use to up- or down-regulate a specific miRNA? Why?
Q6: How do you verify the results of up- or down-regulating an miRNA?
List of Resources

Download the PDF version here

Letter from the Editor

This month, GT brings you a technical guide on manipulating microRNAs. Although the first miRNA was discovered as recently as 1993, scientists have learned much about how these small, non-coding RNAs affect gene expression. Normally, they down-regulate genes by interacting with expressed transcripts, effectively turning them off or dampening their expression. Recent work has shown them to influence at least 30 percent of genes.

Not only have miRNAs been found to play an important role in regulating development as well as protecting against cancer, they are differentially expressed in tissues, making expression profiling a key area of research. And while the toolsets for working with miRNAs — over-expressing them or knocking them out to see what ensues — have greatly improved during the past few years, there are still tricks to master.

To that end, we've consulted some of the best and brightest in the field to get feedback on how best to perform both discovery and expression profiling studies. Their troubleshooting advice ranges from how to isolate and purify miRNA from cells, to how to up- or down-regulate a specific miRNA and how to verify the results. We hope this steers you in the right direction. And for even more thoughtful instruction, check out our resources guide at the back.

— Jeanene Swanson

Index of Experts

Many thanks to our experts for taking the time to contribute to this technical guide, which would not be possible without them.

Galina Gabriely
Brigham and Women's Hospital,
Harvard University

Peng Jin
Emory University

Winston Kuo
Brigham and Women's Hospital,
Harvard University

Joshua Mendell
Johns Hopkins University

Silvia Monticelli
Institute for Research in Biomedicine
Switzerland

Q1: How do you effectively isolate and purify miRNA from cells?

For miRNA isolation we routinely use TRIzol RNA reagent, which is normally used for total RNA isolation. This method is sufficient for purification of miRNAs from cultured cell lines. However, if isolation is required from small or poor-quality tissue samples, we use the mirVana miRNA Isolation kit from Ambion. With this kit, miRNA could be isolated using either total RNA extraction protocol or enriched for small RNAs/miRNA.

— Galina Gabriely

In general we use TRIzol to isolate RNAs from cells. TRIzol has been very effective to recover small RNAs from tissues or cells.

— Peng Jin

In general, several techniques have been tried in our lab, and this depends on the method we would use for profiling. TRIzol has been very effective and the mirVana miRNA Isolation kit from Ambion. Other methods can also be used effectively. For certain technologies, there is no need for isolation or purification; it requires the addition of a lysis reagent. A critical issue with RNA sample preparation is the effective precipitation of small RNA fractions. Traditional methods using 70 percent ethanol washes do not effectively precipitate small RNAs. Increasing the ethanol concentration to >80 percent in these steps will address this problem.

— Winston Kuo

For expression analysis in cells and tissues, straightforward total RNA isolation methods work well. We typically use TRIzol. More sophisticated column-based methods or further purification of small RNA populations has not been necessary in our experience. Furthermore, DNA contamination, which can sometimes be an issue when assessing mRNA abundance in TRIzol-isolated RNA, does not seem to be a problem when measuring miRNAs in these samples. If we wish to specifically analyze small RNA populations in cells (e.g., for sequencing small RNA libraries), we simply gel-purify small RNAs from TRIzol-isolated total RNA.

— Joshua Mendell

If you are not limited with the amount of starting material (i.e., cells or tissues), then a regular TRIzol extraction of total RNA is enough, and you get good quality and reproducible yields. We routinely use RNA purified in this way to detect miRNAs by northern blot and real-time PCR, with very comparable and reproducible results. It is a bit trickier when you have smaller amounts of starting material. You can still just use tRIzol, but you need to improve the RNA precipitation step, and we found that adding some glycogen-blue works just fine. Otherwise you can enrich for the small-RNA fraction using spin-column kits like the mirVana from Ambion.

— Silvia Monticelli

Q2: How do you efficiently clone an miRNA?

It depends on the purpose of cloning an miRNA. If it is for miRNA discovery or high-throughput sequencing, we will use 5'- and 3'-adapter ligation to generate small RNA libraries. If it is a known miRNA we are interested in expressing, we generally use PCR to amplify the DNA fragment containing the miRNA of interest and then verify its expression from an appropriate vector.

— Peng Jin

When cloning an miRNA, there are a couple of issues. If you are interested in cloning a large chunk of the genomic context surrounding the miRNA (e.g., upstream promoter/enhancer regions), this method is often used when endogenous transcriptional regulation and processing (e.g., drosha/dicer action) is under examination. In this instance, standard PCR-based methods can be used to generate miRNA inserts with compatible restriction enzyme sites for your vector system of interest. If you are interested in cloning of a much smaller sequence encoding the miRNA only with little or no flanking sequence, this approach to miRNA cloning is highly efficient. The same protocols that are used to generate shRNA hairpin vectors work well for miRNA cloning. You simply have oligos encoding the miRNA synthesized that contain the appropriate overhangs for cloning and perform an annealing reaction followed by a ligation reaction into the linearized vector. This strategy is often used when straight over-expression is required for the experiment and endogenous transcription/processing is not relevant or even undesirable.

— Winston Kuo

To clone an miRNA-encoding sequence from genomic DNA, we simply PCR-amplify the pre-miRNA hairpin and ~100 bp of 5' and 3' flanking sequence. These fragments can be directly cloned into virtually any mammalian expression vector (with RNA polymerase II promoter) and robust expression will be achieved with rare exception. To directly clone mature miRNA species (e.g., to identify novel miRNAs), small RNAs are first isolated from total RNA by gel purification. After ligating on adapters, small RNA populations are amplified by PCR and then directly sequenced using next-generation sequencing technology (i.e., 454, Soelxa, SOLiD).

— Joshua Mendell

Our lab doesn't really have much direct experience in this. But there are many protocols out there that are widely available and very accurate, and from labs that have done excellent work in the field. The Bartel lab, for example, has posted on its website very detailed protocols to work on miRNAs.

— Silvia Monticelli

Q3: What method do you use to ensure sensitive and specific gene expression profiling of miRNAs? Why?

We usually perform small-scale profiling of small numbers of human miRNAs using multiplex real-time PCR. This method provides better sensitivity and specificity than microarray profiling and therefore is preferable when the number of miRNAs in question is not very high. The accuracy of profiling by multiplexed RT-PCR can be validated by more sensitive, singleplex reaction.

— Galina Gabriely

We generally utilize an miRNA TaqMan assay to determine the expression of specific miRNAs. We have tried multiple approaches and found that both sensitivity and specificity of TaqMan are better. Of course, we also use high-throughput sequencing for digital gene expression profiling of small RNAs, which has good correlation with TaqMan assays.

— Peng Jin

We have used multiple miRNA profiling technologies that include those from Exiqon, Febit, High Throughput Genomics, Invitrogen, and Luminex. Exiqon's microRNA array platform that contains probes with locked nucleic acid-enhanced oligonucleotide capture probes. LNA are modified nucleosides in which the ribose ring is "locked" with a methylene bridge connecting the 2'-O atom and the 4'-C atom. This increases the melting temperature of the duplex 2-8 degrees c per LNA monomer. This vastly improves the thermal stability and specificity of duplexes formed with complimentary miRNA sequences. Incorporation of LNA nucleosides also allows for the generation of tm-normalized capture probes. This is very important for the performance of the microarray, as miRNA are short and have a broad Tm range. Luminex's FlexmiR microRNA panels combine Exiqon's LNA probes to achieve high specificity. HTG, on the other hand, uses complementary nuclease protection probes that are hybridized to the miRNA in the sample, and then an S1 endo/exo nuclease reaction is carried out which destroys mismatched probes in a very sensitive manner.

— Winston Kuo

In my laboratory, we are currently using custom-spotted oligonucleotide microarrays primarily because they are cost-effective and easy to use. For greater sensitivity, there are good commercially-available options such as the Exiqon and Agilent microarrays and the Applied Biosystems TaqMan arrays (which use real-time PCR for profiling).

— Joshua Mendell

After years of northern blots and dot blots to profile microRNA expression, more and more frequently we use the miRNA TaqMan assays from Applied Biosystems, that are extremely specific and sensitive, and the data always correlate very well with our northern blots.

— Silvia Monticelli

Q4: How do you validate your expression results so that you get robust and reproducible data?

We use qRT-PCR with TaqMan microRNA assays and normalize the data on other miRNAs which are known to be stable. One can also use snoRNAs for normalization. In any case, we usually normalize on two to three different molecules to ensure the robust estimation of miRNA amounts. Also, to ensure robust and reproducible data, we perform experiments with biological repeats.

— Galina Gabriely

By comparing different assay formats early on, we have found that miRNA TaqMan assay is very robust and reproducible. For the miRNAs with high abundance, we also confirm the expression data by traditional northern blot.

— Peng Jin

We validate our expression results using three complimentary strategies:
1. qPCR: there are two approaches we have used, the miScript SYBR green system (Qiagen) and TaqMan miRNA assays (ABI). We prefer Qiagen's strategy because it involves a poly-A tailing approach that circumvents the issues presented by the observed endogenous 3'-end miRNA sequence diversity. This diversity is a problem for the TaqMan stem-loop RT-primer strategy. From a technical standpoint a single oligo dT RT reaction (Qiagen) is superior (cost, liquid handling, variability) to the unique stem-loop RT reactions (TaqMan) that must be performed for each individual miRNA of interest. We perform qPCR assays in 384-well plates with four technical replicates/sample and three experimental replicates per treatment group (12 total wells per condition) to get robust and reproducible data. Appropriate no-template and no-primer controls are included.

We perform all the normal steps to optimize any qPCR assay. Pilot experiments are performed to demonstrate the assays are linear (5X dilution series) and to ensure that they yield a single product (melting curve analysis). We also perform an extensive normalization control selection experiment to ensure that 1) the normalization control does not change in our treatment groups and/or cell lines of interest and 2) we can use the delta delta ct relative quantitation method. That is the assays for our miRNA(s) of interest and the normalization control(s) need to have similar (<10 percent difference) slopes as obtained from the 5X dilution series. We start with five controls (U6B snoRNA, snoRNA24, snoRNA49, GAPDH, and select the one to two most appropriate assays).

2. Northern blotting: We use radio-labeled LNA-based oligos as probes for miRNA northern blots. Northerns are run in triplicate (three experimental samples per treatment group) to demonstrate reproducibility. Many miRNAs are found in functional-related families (share 100 percent seed sequence homology and variable 3'-end homology). Therefore, when using qPCR or northerns you must prove that under your conditions the assay only detects your miRNA of interest and not other family members. We do this by spiking synthetic mature miRNAs (IDT) into a corn (Zea) RNA carrier background.
For PCR each synthetic miRNA family member (input) is run against a panel of qPCR assays for each miRNA family member (assay). The value for the correctly paired input x assay combination is set to 100 percent relative detection and mismatched input x assay combinations are expressed relative to 100 percent. Thermocycling conditions can be modified as needed to reduce assay cross-talk with other miRNA family members.

In the case of northerns, synthetic miRNAs (in the carrier RNA) for each of the miRNA family members is run on a 15 percent PAGE gel containing 8m urea. Hybridization temperatures are optimized so that the LNA probe only binds to the miRNA of interest and not the other family members.

Once the appropriate qPCR and/or northern blotting conditions are established using the spike-in strategy they can be applied to experimental samples.

3. In situ hybridization: We also have used Exiqon's DIG-labeled mercury LNA microRNA detection probes to confirm our miRNA expression results and tissue specificity. We developed a working in situ hybridization protocol on our frozen mouse 18.5 d.p.c. embryonic cranial coronal sections.

— Winston Kuo

Careful validation of all microarray results are very important, regardless of the platform used. We generally use either northern blotting (when we have lots of RNA to work with) or real-time PCR (when the RNA is very precious). We have found that both methods produce high-quality expression data when performed properly. Northern blotting is cheaper but requires a lot of RNA, whereas real-time PCR is more expensive but very sensitive.

— Joshua Mendell

I am old-fashioned, I like to see bands on a gel! We make sure that expression data are always consistent using different kinds of approaches. Obviously this is not always possible, as, for example, for northern blot you need 25-30 mg of total RNA per lane
(versus 10 ng for the TaqMan-based approaches) and you just don't always get this much RNA, especially if you are working with primary cells.

— Silvia Monticelli

Q5: What method do you use to up- or down-regulate a specific miRNA? Why?

For down-regulation of miRNA we use antisense oligonucleotides (Aso) modified with
2'-o-metoxyethylribose (moE) provided by our collaborators (Regulus), as they result in the most specific inhibition. In addition, we use commercially available locked nucleic acid (LNA)-modified oligos. Due to their high-affinity binding to miRNAs, LNA-modified oligos provide strong inhibition, though they may have some off-target effects. For up-regulation we use synthetic miRNA mimics available through Ambion.

— Galina Gabriely

We use both RNA oligo- and vector-based approaches. For an RNA oligo-based approach, we apply siRNA duplex-like miRNA and antimir/antagomir to increase or block the activity of a specific miRNA. This approach is very convenient for manipulating the activity of specific miRNA in cell culture. A vector-based approach has the advantages of long-term alteration and the ability to track individual cells. We utilize artificial shRNA plasmids or pol II expression vectors containing a pri-miRNA DNA fragment to over-express a specific miRNA. To down-regulate a specific miRNA, we will either express a shRNA that could produce siRNA against the stem-loop region of an miRNA precursor or use a miRNA sponge that could express a reporter gene with multiple target sites of specific miRNA.

— Peng Jin

There are two general strategies: 1) vector/viral-based over-expression of miRNAs or miRNA antisense sequences and 2) transfection of exogenous miRNA duplexes or antisense inhibitors. The method chosen depends on the design of the experiment. Generally speaking, we favor transfection of exogenous gain-of-function and loss-of-function reagents. Vector/viral-based strategies often rely on sequential processing of the resulting transcripts by drosha/dicer and must be exported from the nucleus via Exportin 5. miRNA processing defects at the level of drosha and dicer have been reported in established human cancer cell lines (the model systems we work in most often). Furthermore, Exportin 5 has been shown to be a bottleneck in the miRNA biogenesis pathway in cell and mouse models. Exogenous over-expression involving DNA integration can saturate Exportin 5 thus competing off endogenous small RNA sequences. In extreme cases this can lead to cell/animal death; however, we presume that less penetrant phenotypes are also possible and could result in serious experimental artifacts. Transfection of exogenous duplexes that enter the pathway via direct incorporation into the RISC complex circumnavigate these pitfalls. Of course, in cases where persistent gain-of-function or loss-of-function is required (e.g., xenograft models of tumor growth) stable expression using vector/viral-based solutions are required. Stable lines must be established carefully, employing viral tittering experiments to avoid the pitfalls described above.

— Winston Kuo

For up-regulating miRNA, we often construct expression vectors (in standard mammalian expression plasmids or viral vectors). Making these is very simple. As described above, we simply amplify the miRNA hairpin and some flanking genomic sequence and clone the PCR products directly =. Alternatively, we sometimes transiently transfect cells with synthetic miRNA mimics (RNA oligonucleotides identical in sequence to the mature miRNA) to over-express an miRNA. With both of these methods (plasmid vs. synthetic mimic), supraphysiologic expression levels are often achieved, so care must be taken in interpreting results from these experiments. To inhibit miRNAs, we generally use commercially available antisense oligonucleotides. We have had success with the inhibitors made by Dharmacon and Exiqon. The limitation of these reagents is that they only work transiently. To achieve stable inhibition, we have started working with so-called "miRNA sponge" constructs, as described first by Phil Sharp's laboratory. These are essentially decoy transcripts with multiple miRNA binding sites that are thought to sequester the active miRNAs in a cell.

— Joshua Mendell

We use mostly lenti- or retrovirus-based transduction systems because we work mainly with primary cells and we need sustained expression for long periods of time. Depending on the cell type, we also use transient transfection with lipofectamine or Amaxa. For down-regulation we are using miRNA sponges. For short-term experiments transient transfection of antagomirs works well, too.

— Silvia Monticelli

Q6: How do you verify the results of up- or down-regulating an miRNA?

To assess miRNA activity following its manipulation, we measure expression of luciferase reporter fused to the perfect binding site of the miRNA. However, the best way to evaluate the activity of miRNA is to assess the expression of its natural targets (when the target of an miRNA is known). This should be done by western blot analysis. Modulation of miRNA expression can be assessed by measuring the expression levels of miRNA using northern blot analysis. In some cases, qRT-PCR with TaqMan microRNA assays may also give reliable results for miRNA expression, but oligos used for miRNA modulation may interfere with primers during PCR reaction, which often affects the real expression data.

— Galina Gabriely

We use several strategies including miRNA qPCR, miRNA northern blot, luciferase reporter assays, mRNA target qPCR, and protein target western blot. I discussed miRNA qPCR and Northern blotting above. These allow for direct detection of miRNA over-expression or down-regulation. Luciferase reporter assays involve cloning of 3'-UTR sequences downstream of the luciferase coding sequence. These 3'-UTRs can contain artificial miRNA cognate sites or endogenous sequences from validated mRNA targets (when this information is available). A critical control in the later strategy is the generation of point mutants in the seed sequence of putative cognate sites. Such mutants should abrogate the ability of the miRNA to regulate luciferase expression.

— Winston Kuo

To verify miRNA up-regulation, we again use northern blotting or real-time PCR to measure miRNA abundance. Measuring miRNA inhibition may not be as simple. Often, antisense oligonucleotides do reduce the abundance of the miRNAs they target. So one can use these same methods to monitor miRNA inhibition. However, sequestering an miRNA in vivo may not always result in a decrease in miRNA abundance. An alternative way to monitor miRNA inhibition is to use a reporter construct with an miRNA binding site. The miRNA will inhibit expression of the reporter unless its activity is blocked. Of course, the use of such reporters is generally limited to a cell culture setting.

— Joshua Mendell

This is a critical step in these kinds of experiments: by transfecting or transducing cells you are effectively somehow changing them, so you really need a lot of controls to make sure that the effect you are seeing is really due to miRNA expression or down-regulation.
It is particularly important with miRNAs because in many cases you don't really expect a strong phenotype, but you are looking for milder effects of 'fine-tuning' gene expression. I think you always need to use at the very least scrambled oligos, miRNAs different from the one you are interested in and mutations in the seed region. I would also recommend using different approaches to try to achieve the same results.

— Silvia Monticelli

List of Resources

Our panel of experts referred to a number of tools that may be able to help you get a handle on working with miRNAs. Whether you're a novice or pro at manipulating microRNAs, these resources are sure to come in handy.

Publications

Chang TC, Yu D, Lee YS, Wentzel EA, Arking DE, West KM, Dang CV, Thomas-Tikhonenko A, Mendell JT. Widespread microRNA repression by Myc contributes to tumorigenesis. Nat Genet. 2008 Jan; 40(1):43-50. Epub 2007 Dec 9.

Johnson SM, Lin SY, Slack FJ. The time of appearance of the C. elegans let-7 microRNA is transcriptionally controlled utilizing a temporal regulatory element in its promoter. Dev Biol. 2003 Jul 15; 259(2):364-79.

Krützfeldt J, Rajewsky N, Braich R, Rajeev KG, Tuschl T, Manoharan M, Stoffel M. Silencing of microRNAs in vivo with 'antagomirs'. Nature. 2005 Dec 1; 438(7068):685-9. Epub 2005 Oct 30.

Lagos-Quintana M, Rauhut R, Yalcin A, Meyer J, Lendeckel W, Tuschl T. Identification of tissue-specific microRNAs from mouse. Curr Biol. 2002 Apr 30; 12(9):735-9.

Landgraf P, Rusu M, Sheridan R, Sewer A, Iovino N, Aravin A, Pfeffer S, Rice A, Kamphorst AO, Landthaler M, Lin C, Socci ND, Hermida L, Fulci V, Chiaretti S, Foà R, Schliwka J, Fuchs U, Novosel A, Müller RU, Schermer B, Bissels U, Inman J, Phan Q, Chien M, Weir Db, Choksi R, De Vita G, Frezzetti D, Trompeter Hi, Hornung V, Teng G, Hartmann G, Palkovits M, Di Lauro R, Wernet P, Macino G, Rogler CE, Nagle JW, Ju J, Papavasiliou FN, Benzing T, Lichter P, Tam W, Brownstein MJ, Bosio A, Borkhardt A, Russo JJ, Sander C, Zavolan M, Tuschl T . A mammalian microRNA expression atlas based on small RNA library sequencing. Cell. 2007 Jun 29; 129(7):1401-14.

Lee RC, Feinbaum RL, Ambros V. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell. 1993 Dec 3; 75(5):843-54.

Lim LP, Lau NC, Garrett-Engele P, Grimson A, Schelter JM, Castle J, Bartel DP, Linsley PS, Johnson JM. Microarray analysis shows that some microRNAs down-regulate large numbers of target mRNAs. Nature. 2005 Feb 17; 433(7027):769-73. Epub 2005 Jan 30.

Marson A, Levine SS, Cole MF, Frampton GM, Brambrink T, Johnstone S, Guenther MG, Johnston WK, Wernig M, Newman J, Calabrese JM, Dennis LM, Volkert TL, Gupta S, Love J, Hannett N, Sharp PA, Bartel DP, Jaenisch R, Young RA. Connecting microRNA genes to the core transcriptional regulatory circuitry of embryonic stem cells. Cell. 2008 Aug 8; 134(3):521-33.

Monticelli S, Ansel KM, Xiao C, Socci ND, Krichevsky AM, Thai TH, Rajewsky N, Marks DS, Sander C, Rajewsky K, Rao A, Kosik KS. MicroRNA profiling of the murine hematopoietic system. Genome Biol. 2005; 6(8):r71. Epub 2005 Aug 1.

Vella Mc, Choi EY, Lin SY, Reinert K, Slack FJ. The C. elegans microRNA let-7 binds to imperfect let-7 complementary sites from the lin-41 3'UTR. Genes Dev 2004 Jan 15; 18(2):132-7. Epub 2004 Jan 16.

Websites

miRBase
http://microrna.sanger.ac.uk/sequences

microRNA.org
http://www.microrna.org/microrna/home.do

Target Gene Prediction at EMBL
http://www.russell.embl.de/miRNAs

Conferences

MicroRNA and Cancer
Keystone Symposia on Molecular and Cellular Biology
MicroRNA in Human Disease & Development
Cambridge Healthtech Institute