TB #152: Choice of Ribonucleases for a Ribonuclease Protection Assay

Ribonuclease protection assays (RPAs) are a technique used for detection and quantitation of specific RNAs. The method is based on the ability of ribonuclease to specifically degrade single-stranded RNA while leaving intact a labeled antisense RNA probe that is hybridized to its target. Ribonuclease protection assays are more sensitive than Northern blots for the detection of mRNAs, more tolerant of partially degraded RNA, and are able to distinguish between transcripts of multi-gene families that co-migrate on Northern blots. Using a high specific-activity probe (approximately 10⁹ cpm/µg), a moderately abundant mRNA such as ß-actin can easily be detected in sub-microgram quantities of total RNA.

The original published RPA procedure used RNase A, which cleaves after C and U residues (Melton, P.A. et al., 1984). Subsequently, mixtures of RNase A and RNase T1 were used (Winter, E. et al., 1985). Since RNase T1 cleaves after G residues, the degradation of single-stranded probe was assumed to be more complete when using the RNase A/T1 mixture compared to using either RNase alone. RNase T2, which cleaves after all 4 residues, but preferentially after As, has also occasionally been used for ribonuclease protection assays. Recently, RNase I from E. coli has been cloned and become commercially available. RNase I cleaves after all 4 ribonucleotides with no base preference and might therefore be expected to be well-suited for use in RPAs. We have compared various RNases in standard ribonuclease protection assay procedures, and have also investigated their use in specialized procedures such as single-base mismatch detection and protection of A-U-rich hybrids. (Ambion's MutationScreener™ Kit is based on RNase digestion for single-base mismatch detection.) The following is a discussion of our results.

A frequently asked question is whether a particular RNase offers advantages in terms of sensitivity in ribonuclease protection assays. To answer this, we compared RNase I with RNase A/T1 mix across a wide range of RNase concentrations, using a probe for mouse -actin mRNA hybridized to 5 ľg of total mouse liver RNA. The full-length 300 nt probe contains 250 nt of sequence that is complementary to -actin mRNA. Figure 1 shows that for both types of RNases, sensitivity is a function of RNase concentration. Digestion with suboptimal amounts of RNase results in failure to degrade all single-stranded probe; this is shown in the first lanes of both panels, where the expected size shift from 300 nt to 250 nt (representing the difference between full-length transcript and protected fragment) is incomplete. In contrast, the 7th lane of the 1st panel, and lanes 5-8 of the 2nd panel show the effect of using too much RNase A/T1 mix or RNase I, respectively, in the RPA. Over-digestion with RNase causes partial degradation of the protected probe, resulting in a loss of sensitivity. It is apparent that both the RNase A/T1 mixture and RNase I are suitable for use in the RPA. However, the RNase A/T1 mix shows a wider concentration range over which it is effective for specifically degrading unhybridized probe. Thus, for standard applications, the mixture of RNase A/T1 is the best choice for ribonuclease protection assays. There are, however, some circumstances where other ribonucleases may offer advantages.

Figure 1. Effect of Ribonuclease Type and Concentration on Signal Strength - Comparison of RNase I and RNase A/T1 Mixture in an RNase Protection Assay. Total mouse liver RNA (5 µg) was hybridized overnight at 42°C to 10⁵ cpm of ³²P-labeled probe (300 nt mouse ß-actin antisense transcript, 1.3 x 10⁹ cpm/µg) in 20 µl of RPA II™ Hybridization Solution. Reactions were treated with 200 µl of RNase Digestion Buffer containing the indicated amounts of RNase for 30 minutes at 37°C. Ribonuclease was inactivated and protected RNA precipitated by following the RPA II protocol. Half of each sample was analyzed by 5% PAGE and autoradiography. The control lane shows the undigested full-length probe (no RNase added). NOTE: One unit RNase A is equal to approximately 2 µg. Note also, unit definitions differ between the RNases.

If the probe-target pair in a ribonuclease protection assay is A-U rich, problems obtaining full-length protected fragments are frequently encountered. This is often a problem with 3' untranslated region probes since they tend to be A-U rich. Under standard RNase A/T1 digestion conditions, there is apparently sufficient transient strand separation ("breathing") to allow the ribonucleases to nick the duplex RNA. Transient strand separation can be reduced by stabilizing the duplex either by lowering the temperature or increasing the salt concentration of the digestion buffer. However, both of these approaches are inconvenient, and high salt reduces the activity of RNases. The simplest approach is to use RNase T1 by itself instead of the RNase A/RNase T1 mixture. Since the RNase T1 only cleaves after G residues, the A-U rich regions most susceptible to breathing will be refractory to cleavage by RNase T1. This generally solves the problem, although some probe-target combinations give higher background signal than with the RNase A/T1 mix.

A comparison of different RNases in a ribonuclease protection assay with an A-U rich hybrid is illustrated in Figure 2. The 550 nt probe contains 505 nt of sequence complementary to the 3' end of a sea urchin cyclin B gene. The protected fragment is 72% A + U. Only digestion with RNase T1 alone results in protection of the full-length fragment; digestion with RNase A or RNase I alone, or with the RNase A/T1 mixture, gives an array of smaller bands, with very little full-length protected probe. The control lanes show that digestion with each RNase alone is sufficient to completely degrade unhybridized probe. Note the small amount of low molecular weight "background" fragments seen in the RNase T1 control lane. RNase T1 alone may also be used to obtain a single discrete product when there is a sequence divergence between the probe and target. For example, the mouse ß-actin control probe supplied with the Ambion RPA II Kit has 11 single-base mismatches when hybridized to rat ß-actin mRNA. When the mouse probe is used with rat RNA, cleavage at probe/target mismatched positions is minimized by using RNase T1 alone.

Figure 2. Comparison of RNases in a Ribonuclease Protection Assay with A-U Rich Hybrid. Total sea urchin RNA (9 µg) was hybridized to 5 x 10⁴ cpm of ³²P-labeled probe (a 550 nt sea urchin cyclin B 3' untranslated region antisense transcript, 6 x 10⁸ cpm per µg) in 20 µl of Soln A (hybridization buffer containing 80% formamide) and incubated overnight at 42°C. Reactions were treated with 200 µl of RNase Digestion Buffer containing the indicated amounts of RNase for 30 minutes at 37°C. RNases were inactivated and protected probe precipitated by addition of 300 µl of Inactivation/Precipitation Solution. Pellets were resuspended in 8 µl of gel loading buffer, heated 3 minutes at 95°C, and analyzed by 5% PAGE and autoradiography. The control lanes contained probe plus 10 µg of yeast RNA digested with RNase as indicated or incubated with RNase Digestion Buffer only. The full-length protected fragment is 503 nt, consisting of 72% A + U. Note, unit definitions differ between the RNases.

The ability of RNase I to efficiently cleave after all 4 bases may offer advantages in performing ribonuclease protection assays under some circumstances. It is probably the enzyme of choice for distinguishing between members of closely related gene families or even allelic variants. Regions of single-base mismatch vary in their ability to be cleaved by both RNase A and RNase I. We have examined two types of mismatches, an A/C and a G/U, which in the literature are reported to be relatively easy and difficult, respectively, to cleave with RNase A. The easy-to-detect A/C mismatch is also efficiently cleaved by RNase I, while the difficult-to-detect G/U mismatch is not cleaved at all by RNase I. However, the A/C mismatch yielded a cleaner result with RNase I. Over-digestion with RNase I results in partial degradation and smearing of the subfragments to a smaller size, whereas over-digestion with RNase A tends to generate discrete non-specific fragments in addition to the expected subfragments. At this point, it appears that both RNase I and RNase A offer different advantages and disadvantages in detecting single-base mismatches. Further experiments are in progress to define optimum conditions for mismatch detection with these two enzymes. Regions of mismatch larger than single bases are cleaved much more efficiently than single-base mismatches. We predict that these regions will in general be cleaved more efficiently by RNase I than RNase A due to RNase I's ability to cleave after all 4 bases. RNase I may also to be superior to RNase A or A/T1 mix for mapping studies where it is desirable to have single-base resolution, since RNase I should be able to trim probe flush to the target irrespective of the sequence. Although RNase T2 theoretically offers the same advantages as RNase I in being able to cleave after all 4 bases, it has not been widely used in ribonuclease protection assays, probably because it is expensive and because it shows a strong preference for cleavage after A residues (Uchider and Egami, 1967).