User:ElNando888/Blog/A minor glitch?

I recently presented the Adenine as a nucleotide a little out of the ordinary, an "outsider" of sorts. Well, there's more to it...

Motif

Adenine is known to form motifs known as "A-motif". Apparently, the nucleotide fits beautifully on the minor groove side of the GC pair. In contrast, the pyrimidines are too small to match well, and the Guanine has a disturbing bulky NH group on its Watson-Crick edge, that makes things difficult locally.

I'd like to invite the readers to take the time (trust me, it will be well spent) to read the following scientific papers (don't worry, these are not as technical as many others). And for each of those, I quoted the ideas that seemed important, at least to me.

A general paper about RNA motifs: http://dasher.wustl.edu/bio5357/reading/accchemres-44-xxx-11.pdf

"A Universal Packing Strategy for RNA"

A more in-depth paper completely focused on A-motifs: http://www.pnas.org/content/98/9/4899.full

"A-Minor Motifs Are Common in All Large RNAs"

"A-Minor Motifs Are More Abundant than Tertiary Base Pairs"

"A-Minor Motifs Often Cluster"

http://www.pnas.org/content/98/9/4899/F1.expansion.html

For the sake of illustrating how A-motifs cluster, an example from the PDB accession 3CC2

Types I, II & III in what they call an "A patch".

Flashback

Let's return for a moment to the results of the recent Semicircle 2 bends lab.

Our designs are first "denatured" (another way to put it is to say that they are melted, but it isn't a proper way to describe the procedure), by bringing them to a temperature of 90°C. In the case of this specific design, it is actually questionable whether the main long stacks actually break apart even at that temperature. In any case, they certainly nucleate (form a helix) at very high temperatures. The barcode hairpin though, probably cannot nucleate before the temperature goes down a lot more. So, for a very long time (specially at atomic scales), the secondary structure looks like the picture above.

Now let's review:

We have long and very stable stacks, exclusively composed of canonical base pairs, making the helix as rigid as it can be, and with a fair number of GC pairs.
And we have a bunch of Adenine doublets freely roaming...

Maybe, preventing a stack from forming is as simple as making sure that the nucleotides are far away from each other... What I'm trying to say is that I suspect the primer-complementary tail, with all its A-doublets, to interact with the already formed long stacks (not the barcode region), dragging one side (the 3' one) of the barcode with it, while the other side of the barcode (the 5' one) is of course attached to another location. The barcode wouldn't form, because the bases would just be kept too far away from each other.

In any case, the suspected A-minor interactions would be "SHAPE-invisible". The bases involved are either paired, in which case they will appear even more so SHAPE-protected, and we have no data about the Adenines themselves.

In this quick look at the lab results of the Semicircle lab, I mentioned about some exceptions to the barcode "debacle". Turns out that GU pairs or noncanonical pairs:

do not provide a potential "grip" for an A-motif type I
and do make the helix that contains them more flexible locally, creating instabilities and variations in the intervals between minor grooves

So, these recent strange barcodes... a minor glitch? or possibly, an A-minor glitch?

Testing theories

Isn't it what science is all about? Theories are just worthless if they are untestable. Here though, as I hinted above, it looks like a tough nut. We can't get SHAPE data about the tail, and even if these interactions do occur, they don't seem change anything in the SHAPE signal for the canonically paired bases they are supposed to bind to. So, what kind of lab experiment could validate or falsify the hypothesis?

Current labs

So far, the only vague idea that I got is like following.

If I had the chance, I'd like to rerun janetmason's design, along with a nearly perfect clone, having only 2 pairs swapped in the barcode hairpin, like depicted above.

The predictions for those 2 sequences, by ViennaRNA 2.1.x

Original

Results for thermodynamic ensemble prediction
The free energy of the thermodynamic ensemble is -55.20 kcal/mol.
The frequency of the MFE structure in the ensemble is 85.22 %.
The ensemble diversity is 0.30 .
You may look at the dot plot containing the base pair probabilities [EPS|PDF|IMAGE CONVERTER].
The centroid secondary structure in dot-bracket notation with a minimum free energy of -55.10 kcal/mol is given below.
[color by base-pairing probability | color by positional entropy | no coloring]
1 GGAAAAGCUAGUAUGUACGAGACGUAGUAUGUAGGGAGACCUACAUACUACGCCGUACAUACUAGCAAGCUUCUCUUCGGAGAAGCAAAAGAAACAACAACAACAAC
1 ......(((((((((((((.(.(((((((((((((....)))))))))))))))))))))))))))..(((((((....))))))).....................

Mutated

Results for thermodynamic ensemble prediction
The free energy of the thermodynamic ensemble is -55.00 kcal/mol.
The frequency of the MFE structure in the ensemble is 85.04 %.
The ensemble diversity is 0.32 .You may look at the dot plot containing the base pair probabilities [EPS|PDF|IMAGE CONVERTER].
The centroid secondary structure in dot-bracket notation with a minimum free energy of -54.90 kcal/mol is given below.
[color by base-pairing probability | color by positional entropy | no coloring]
1 GGAAAAGCUAGUAUGUACGAGACGUAGUAUGUAGGGAGACCUACAUACUACGCCGUACAUACUAGCAAGGUUGUCUUCGGACAACCAAAAGAAACAACAACAACAAC
1 ......(((((((((((((.(.(((((((((((((....)))))))))))))))))))))))))))..(((((((....))))))).....................

ViennaRNA thought the original was a sure win. As we could observe, things didn't go as planned. My proposed mutation is extremely close in predicted numbers. In all logic, these two designs should either succeed or fail the same way. So what am I hoping to see here?

Do you guys remember Jennifer Pearl and her work on dotplots? At the time, I wouldn't follow too closely that work, because I always had doubts about the accuracy of the dotplot in general (and we tested recently that for instance, pseudoknots are totally invisible for ViennaRNA). This said, a number of things she was chasing after was making totally sense to me. Specifically, the possibility that certain misfolds may have an actual positive effect on the final result. The idea being basically that the weak and temporary misfolds prevent the sequence from forming more dangerous ones and thus, prepare the free energy landscape in a way that funnels the vast majority of the ensemble to the targeted secondary structure, usually the MFE. (She was also studying how these intermediary structures may have been a helpful device for conformational changes in the context of riboswitches, but that's another story)

In this case, I'm simply hoping that this misfold may help "sequester" the tail and prevent it (at least a little) from interacting with the rest of the design. And at the same time, this misfold shouldn't be strong enough (here is one serious difficulty, -5 kcal, is that too strong already? or too weak to have any effect?) to hold for too long, hopefully allowing for the barcode hairpin to form properly and rapidly at a later point.

Alas, it may not be possible to reserve barcodes for this purpose...

A lab proposal

(coming soon)