There are a number of strategies that have been proposed for cloning small RNAs. Before discussing these, however, there is one factor common to all of them that is essential to be aware of. Small RNAs, whether from plant cells, animal cells, or other sources, represent a small fraction of the total RNA mass present. Agilent Technologies quantifies the quality of cellular RNA in the form of their RNA Integrity Number (RIN). Very high quality intact RNA has a RIN of 10.0 and the lower the RIN, the more degraded the RNA. RIN values between 6.5 and 10.0 represent a continuum of acceptable to excellent RNAs. Using RIN as the point of departure, Agilent assessed the relative fraction of total RNA that is within the small RNA size range in forty tissues from human, mouse, and rat [18
The results, summarized in , show two important features. First, for all but five tissues, the relative mass of small RNAs is below 3% and, second, there is a significant negative correlation (r = −0.58; P < .01, df = 38) between overall RNA quality as assessed by RIN value and relative small RNA mass. Clearly, increasing amounts of RNA degradation will introduce a greater mass of small fragments that lie in the true small RNA zone. This will result in a greater mass of competing RNA that will make it more and more difficult to see the real small RNAs that are the targets of interest even if the majority of the degraded RNAs are themselves unclonable by some of the methods discussed below. While there will be variation from RNA source to RNA source, it is clear that larger RNA components like mRNAs, rRNAs, and tRNAs, comprise by far the bulk of the total RNA and that the relative mass of the true small RNA fraction should and will be the smallest in very high quality RNA. A generalized RNA mass profile for high RIN RNA is presented in . As can be seen, the true miRNA region is indeed a very small part of the total mass. Given this, it is essential to the small RNA cloning process that RNA quality, as assessed by measures like RIN, be as high as possible and that as much of the competing RNA mass as possible be removed so that a “target-rich” small RNA component can be purified prior to starting the cloning process.
Figure 1 Linear regression of total RNA quality (RIN) and the relative mass of the small RNA population determined for forty human, mouse and rat tissues. A significant negative correlation coefficient, r = −0.58, P < .01, df = 38, derived from (more ...)
Figure 2 Mass profile of human RNA. Here, the absolute mass fractions of RNAs up to 4000 nt in length are shown. The position and composition of the small RNA region, defined as that portion of the total RNA mass that is between 0 and 200 nt long are highlighted. (more ...)
Small RNA enrichment can be accomplished in a number of ways. One of the simplest ways is to simply run a sample of total RNA on a denaturing polyacrylamide gel (dPAGE) and excise the area of the gel containing the small RNA fraction (see the appendix). The problem with this method is that the enriched small RNAs must be removed from the gel and purified for further manipulations and this routinely results in a substantial loss of what is already a small amount of mass to begin with. There are ways to minimize this loss of material and we will discuss one of these in the next section. Other methods for enriching the small RNA fraction have been developed including column capture and release methods like the mirVana protocol from Ambion and the timed size exclusion method, represented by the flashPAGE fractionator system, also from Ambion. The point is that, whatever method is employed, the small RNA fraction of total cellular RNA must be enriched to increase the likelihood of successfully cloning small RNAs.
Once the small RNA fraction is enriched and purified, there are several ways to proceed to clone the individual small RNAs contained in the fraction. Berezikov et al. [19
] reviewed the basic small RNA cloning methods. In all cases the target species for direct cloning is an RNA varying in size between 18 and 25 nucleotides (nt) having a free 3′ hydroxyl group and a free 5′ phosphate group. Although some variation exists [20
], the universal initial step in the cloning process is first to ligate a 3′ adaptor sequence through the free 3′ hydroxyl. The 3′ adaptor will serve as the site for later annealing of an oligonucleotide primer for reverse transcription. As seen in , there are several possible ways to accomplish this adaptor joining. In one option, the small RNA species are polyadenylated creating a 3′ extension [21
]. However, as many small RNA species in plants have been shown to contain 2′-O-methyl modifications on their 3′ ends, this method may be of only limited utility since such modifications block polyA polymerase extension [22
]. Both of the other 3′ adaptor joining options are designed to prevent later circularization of the linkered RNAs. In one variation, the RNAs are dephosphorylated prior to adaptor ligation and then rephosphorylated for subsequent processing [23
]. In the other variation, the 5′ end of the adaptor is preadenylated and the 3′ end blocked by a nonstandard group such as a dideoxynucleotide [10
]. Preadenylation of the adaptor obviates the need to dephosphorylate the target RNAs because the adaptor joining via T4 RNA Ligase can be carried out in the absence of ATP. Given the obvious advantage that this method confers by reducing the number of operations required to process target RNAs, New England BioLabs (NEB) has introduced a truncated T4 RNA Ligase that specifically reacts with preadenylated 3′ linkers [25
]. Regardless of the method chosen, however, producing a stable and reactive 3′ linkered small RNA population is the goal of the first step in cloning.
Figure 3 Diagram of extant small RNA cloning strategies. Following small RNA enrichment, all strategies share the same outline of first placing an adaptor on the 3′ end of the target RNAs, then placing a second adaptor on the 5′ end of the RNAs, (more ...)
The next phase of cloning is to join a second adaptor to the small RNA population. This time, the adaptor is joined to the 5′ end. As shown in , there are now but two ways to do this and the choice is dictated by the methods chosen for 3′ adaptor joining. If the method chosen is the polyadenylation route, then the 5′ adaptor joining method is to carry out a template switch. This method relies on the property of a number of reverse transcriptases to add a small number of nontemplated nucleotides to the 3′ ends of cDNAs. Since the nontemplated nucleotides tend to be mostly deoxycytidines, an adaptor containing a poly-G 3′ run can be used to switch the template from the miRNA to the adaptor [19
]. The other path is to use a 5′ adaptor with a 3′ hydroxyl group that will ligate to the 5′ phosphate of the target RNAs. This is carried out with a T4 RNA Ligase in the presence of ATP and is followed by a reverse transcription using a primer complementary to the 3′ linker. In both cases, the resulting cDNA population is PCR amplified in preparation for cloning and/or sequencing.
PCR amplicons can be directly cloned using any one of several PCR cloning vectors or the amplicons can be processed to form concatamers which are then cloned. Concatamer formation from amplicons is a direct descendant of the Serial Analysis of Gene Expression (SAGE) methodology developed in the 1990s by Velculscu and colleagues [24
]. The obvious advantage of concatamer cloning is that individual clones will contain more small RNAs than the ones that will be present if the PCR amplicons are simply shot-gun cloned. This is a consideration for conventional Sanger dye-terminator sequencing but, as will be discussed later, new generation deep sequencing methods have circumvented the need for concatamers and, indeed, for cloning at all.
One aspect of the cloning methods shown in is that small RNAs will all contain a 5′ phosphate group following 3′ adaptor joining. This constant feature that allows for subsequent 5′ adaptor joining was believed to represent the universal state of small RNAs in vivo. In 2007, Pak and Fire [29
] announced that this is not the case. Attempts to clone a specific small RNA in C. elegans
called Cel-1 repeatedly failed even though there was ample evidence that it existed. Their persistence in uncovering the reason for Cel-1 being refractory to conventional small RNA cloning methods paid off in their discovery that Cel-1, and, now, other small interfering RNAs, was tri-phosphorylated on its 5′ end [29
]. They developed an alternative method for cloning troublesome RNAs featuring the use of two
3′ ligations with the reverse transcription step in between the two ligations. This alternative method, named by them 5′ Ligation Independent Cloning, is completely indifferent to the state of the 5′ end of the target RNAs. The reverse transcription step following the initial 3′ adaptor ligation makes the initial 5′ end the new 3′ end with a hydroxyl group ready for a second 3′ ligation step regardless of what may or may not have been present on that initial 5′ end. The 5′ Ligation Independent Cloning option revealed that a secondary pool of small RNAs was being produced in C. elegans
via a completely different pathway from conventional miRNAs [29