In Situ Hibridization of Very Small Sequences - Northern Blot of miRNAs (Jan/25/2006 )
Hello there:
I'm trying to perform an ISH of miRNAs, small sequences of 17-24 nt, on slices from different rat tissues. The issue is that i need to have real specificity, given the fact that any 21 nt sequence may be found in almost every RNA. So if I am to do it, i'll need to decrease the noise/signal ratio.
So far i've thought of these:
1) To sinthesize random DNA oligos and pre-treat the samples with them, in order to hibridize almost any RNA. Then perform the ISH with my specific RNA probes. Of course many signal-positive molecules will be "titrated" by the DNA oligos but, of course, any signal detected by the RNA probes will be reliable.
2) To use small amounts of Salmon Sperm DNA during the pre-hibridization stage to achieve the same goal.
3) A combination of the previous two.
I've dug on the literature and found only two papers that use this technique and they seem to use Salmon Sperm DNA, but to be honest im afraid that i could have false-positive signal.
My final goal is to perform it in ultrastructural studies.
So the question is, what would you do in my case?. I want it all, specificity and a strong signal. By the way I'm using DIG as a tracer.
I'm trying to perform an ISH of miRNAs, small sequences of 17-24 nt, on slices from different rat tissues. The issue is that i need to have real specificity, given the fact that any 21 nt sequence may be found in almost every RNA. So if I am to do it, i'll need to decrease the noise/signal ratio.
To be honest I do not think that a 21nt sequence will occur in almost every, or even many RNA sequences, barring of course conserved sequences which will occur more often. If you have a 21 base sequence, each position having 4 possibilities you only have a 1/4^21 chance of it occuring. Unless my math is wrong, this is a one in 4,398,046,511,104 chance of occuring randomly (again barring conserved sequences). One of the reasons RNA interference is such an elegant answer to viral infections ect... is that the 21nt sequences are usually very specific against the intruding RNA. Granted if you use a sequence to similar to another sequence you may get silencing or reduction of the other gene but if that is the case, change to a different part of the sequence.
The few times I have done ISH (I am no expert in this field) we used salmon sperm DNA to lower background as well.
I agree with Captain DNA, It is very rare case for a 21 sequences found exactly matched sequence on other RNAs.
As to probe, you might thought about Dig labeled LNA probe, which can be purchased from Exiqon Inc. and they also provide some protocols for in situ hybridization on web site.
My experience: LNA modification indeed increasing hybridization signal, but for fear of unspecific signal due to increased affinity of LNA modified probe, you could check it by northern hybridization with such such probe in total RNA to make sure the hybridization signal come from desired size small RNA...
Following is a recent paper I found for you reference:
RNA. 2006 Feb;12(2):187-91. Epub 2005 Dec 22.
Good luck
Rui