Need help with base stacking interactions in ssDNA/dsDNA

[MirkoZlikowski]

Dear forum members,

I have a problem with understanding how base-stacking interactions contribute to the stability of dsDNA. My lack of understanding is probably caused by my insufficient knowledge of basic chemistry and I was wondering if anyone here can explain, in relatively layman's terms, what base stacking is and how it contributes to the stability of double-stranded DNA. The question is a bit longer because I would like to avoid getting answers that contain the type of information that prompted me to pose this question in the first place. So, here we go with my stupidity leak.

In various resources (books, on-line sites, etc...) which describe the types of interactions that keep the two strands of DNA together, usually three types of forces are mentioned: hydrogen bonding, base stacking and hydrophobic effects. Many people, including older ones like myself, have been taught that the two DNA strands are held together by hydrogen bonds between complementary A-T and G-C base pairs (2 bonds per A-T base pair and three per G-C base pair). However, not-so recent articles (such as "Base-stacking and base-pairing contributions into thermal stability of the DNA double helix", Nucleic Acids Res (2006) 34 (2): 564-574.), clearly demonstrate that "base-stacking is the main stabilizing factor in the DNA double helix", whereas hydrogen bonds play minor, if any role. The problem is that, after consulting several textbooks, forums, articles, etc, I still don't understand what base stacking is and how it works. But first let's start with hydrogen bonding:

The story I was taught goes like this: two complementary ssDNA strands, instead of remaining separated, tend to form dsDNA because in dsDNA multiple hydrogen bonds between complementary base-pairs will be formed. However, when DNA is in single stranded form, bases interact with water (in aqueous solutions) and form an even greater number of hydrogen bonds with water molecules than they do when they are paired with complementary bases in dsDNA form. Therefore, formation of dsDNA from ssDNA cannot be driven by the creation of (greater number of) hydrogen bonds between base pairs, since many more hydrogen bonds between water molecules and bases in ssDNA need to be broken to form the double helix in the first place. Therefore, hydrogen bonds are not the drivers behind dsDNA formation and do not play a major role in holding the two DNA strands together - i.e. they provide base-pairing specificity for replication/transcription/recombination, but do not contribute much to the DNA stability. Is this correct?

Enter stacking and hydrophobic interactions. If I understand correctly, since bases are hydrophobic, in aqueous solutions they will tend to align one over the other in order to minimize their hydrophobic surface area that is in contact with water. Bases achieve this by aligning their rings into a parallel orientation, i.e. one base more or less "sits", or stacks, on top of another (for simplicity not going into twist and roll angles, slide etc). Anyway, I think I can understand why this happens at a very primitive level - if we have two bases separated by some distance in water, there will be a cage of water molecules around them. However, if two bases stack on top of each other, water will be released from between the two stacked surfaces, which will release water molecules from an ordered system (cage) into a more disordered system (free water). So, even though we have more ordered system with respect to bases, much more water molecules become disordered, total entropy of the system is larger and base stacking is therefore entropically favored. Is this correct?

So, if what I wrote is at any level correct, I can very loosely 'define' base stacking as an ordered arrangement of DNA bases one on top of the other, which is favored by mysterious factors such as entropy/hydrophobicity.

Given all this, my questions then are as follows:
1) Will single-stranded DNA form stacked structures in aqueous solutions (apart from polyA, which is known to do so? In other words, do base-stacking interactions form only in dsDNA, or also in ssDNA? If they occur only in dsDNA, that can mean that they are not as energetically favored in ssDNA, as they are in dsDNA. To put it in other words, does ssDNA by itself adopt analogous (same?) helical structure as if it would when paired with complementary DNA strand? If not, then why not?
2) How stacking stabilizes dsDNA? Let's take a 2bp dsDNA molecule as an example:

5'3'
| |
A-T
G-C
| |
3'5'

Here A is stacked on top of G and T is stacked on top of C. How does the AG stacking and TC stacking hold the two DNA strands together? I can accept that AG stacking keeps A and G on top of each other, but how does it keep them paired with the complementary strand (TC), i.e. what keeps the two strands locked together if not hydrogen bonds? Does A interacts with C via some sort of cross-stacking interactions? And are then those cross-stacking interactions what holds the two single DNA strands together?

3) If we mix two non-complementary ssDNA molecules in water, due to their hydrophobicity bases in those ssDNAs should also tend to "avoid" water. Will those ssDNAs then form some sort of weird structure in which bases are going to be stacked in the middle of the structure, but not form hydrogen bonds (since they're not complementary?). Or, are they going to remain single-stranded and unstacked. If latter is true (which I don't know), then how can we claim that hydrogen bonds do not play a major role in stability and formation of dsDNA?

So, thanks everyone for reading this and apologies for the tremendous amounts of ignorance I spat out in this post. I hope someone will be able to clear things up for me. I'm offering a pizza and a beer, this has been bugging me for too long.

[BabcockHall]

This is a very big topic, too big to do full justice. If you have not consulted a good biochemistry textbook or two, that would help. Here are a few quick thoughts. One, hydrogen bonding is important, owing to the chelate effect. To satisfy the same hydrogen bonds with water as are satisfied with complementary DNA, a large number of water molecules would have to lose translational entropy. Two, according to one textbook, base stacking is driven by enthalpy and opposed by entropy. Therefore, it is unlike a textbook hydrophobic effect. Unlike hydrogen bonding, base stacking is not a highly cooperative interaction. I have heard that both H-bonding and base stacking make about the same magnitude of contribution to the stability of the double helix.

 

My understanding is that there is base stacking in single stranded nucleic acids, but that the process of stacking is not highly cooperative, unlike H-bonding.

Edited March 28, 2017 by BabcockHall

[BabcockHall]

One more thought: Given that the percentage of GC base pairs is positively related to the T(m) of duplex DNA, hydrogen bonds would seem to have some role.

[CharonY]

Not necessarily. As long as G/C-stacking contribute more to stability, the effect would be the same. One should note that the model has only been tested on relatively short DNA stretehes. But at least in that regime it appears that the contribution of base-pairing has much less influence than previously thought (most Tm calculators use both aspects).

[BabcockHall]

A fair point, and perhaps I should have added that in the presence 5.2 M betaine, the % composition of AT vs GC doesn't affect Tm: PH Von Hippel and collaborators Biochemistry 1993 32, p. 137.

[BabcockHall]

Not necessarily. As long as G/C-stacking contribute more to stability, the effect would be the same. One should note that the model has only been tested on relatively short DNA stretehes. But at least in that regime it appears that the contribution of base-pairing has much less influence than previously thought (most Tm calculators use both aspects).

 

There is a table in Miesfield and McEvoy's biochemistry textbook that indicates that certain base stacking interactions among G:C base pairs are much stronger than the same interaction in certain A:T base pairs. This subject is more complex than I thought it was.

[CharonY]

Yes, in many cases the rough models do not seem to make much of a difference in practical terms, but once you get into details it is quite complicated and I run out of expertise rather quickly to properly evaluate the data. It really becomes more a physics/physical chemistry question.

 

In most molecular classes we tend to teach the abridged version as it is far more intuitive to understand and most of the time it gets us reasonably close to where we want to be. But at night I do wonder about how much we get wrong and how relevant it may (or may not be) for biological questions.

[Velocity_Boy]

Base stacking is just the way of getting the structural integrity of the DNA double helix molecule.

 

Its just A section of DNA. You already know that The bases lie horizontally between the two spiraling strands. So.... The DNA double helix is stabilized mainly by two forces: hydrogen bonds between nucleotides and base-stacking interactions among aromatic nucleobases. That's it, amigo!

[CharonY]

Yes indeed, if we are not interested in the actual details such as relative contributions of the forces, media influence, complex conformational changes that occur in vivo (and under influence of proteins), then it becomes almost trivial. Who knew?

[BabcockHall]

Good points. I would also like to add that the cooperativity of hydrogen bonding and base stacking as something that is worth pondering.

[CharonY]

Yeah, it is difficult to distinguish these two forces experimentally, without a good model and the question is often how precise measurements have to get to be able to accurately reflect the respective contributions (and even more difficult to figure out whether they are biologically relevant). The model proposed in OP does indicate that melting behaviour can be explained by base stacking alone. Theoretically so do base pairing models, though they are less accurate when they do not take a least a fudge factor from the stacking into account..