Temperature_issue

2 replies [Last post]
LC
Offline
Joined: 02/15/2012

Dear Prof. Zuker,

I would like to kindly ask you for your opinion on the following issue:

We are currently dealing with the Second Internal Transcribed Spacer (ITS2) of algae.
We have been folding Helix1 at 37°C in Mfold (version with fixed temperature) and getting two outputs - 1) a branched helix and 2) an unbranched. The free energies of both foldings are almost identical 1) dG = -26.40 [Initially -27.50] and 2) dG = -26.60, respectively.
If the temperature is changed at 47°C (version 2.3 energies) there is only one output - the unbranched helix (dG = -17.48). If the same sequence is folded already at 32°C (version 2.3 energies) the only output is the branched helix (dG = -25.28, no initial free energy).

So far we have been using solely 37°C for folding. In order to find a consensus of ITS2 for a group of taxa, sequential comparison of each bp or nt in the sequence among all taxa has been done.

Now I am just thinking, what is the conclusion? Should be one result (folding/s) considered as more optimal than the others? On one hand our organism are actually living at 17 °C (an average), so theoretically it would be more optimal to use lower temperature. On the other hand according my knowledge most experimental studies of RNA thermodynamics have been performed at or near 37°C. Therefore, I am just wandering about the accuracy of folding at e.g. 17 or 50 °C.
Or it is possible that highly conserved regions of helices show the comparable structure at wide spectrum of temperature?
Can you please give us any suggestion how we can deal with situation and how we can discuss it?

Thank you very much,
LC

LC
Offline
Joined: 02/15/2012
Temperature_issue

Thank you very much for your valuable comments! I have sent the requested sequence to your email address. Best regards, LC

zukerm
Offline
Joined: 11/12/2010
You have asked an important

You have asked an important question. Leishmania lives at 25° C in the sand fly. When it infects a mammalian host, the temperature shifts to 37° C, and a heat shock protein is expressed. It is thought that the signal is a change in secondary structure in the 3'-UTR of the corresponding mRNA. Computer simulations show a change in optical density as temperature rises. These simulations match observations. None of this proves anything, but these results are compatible with the notion of different structures at different temperatures.

Are the predicted foldings well-defined? That is, how cluttered are the energy dot plots? If the energy dot plots are uncluttered, the folding predictions are more credible. I have new tools using the UNAFold package that compute the entropy of the Boltzmann distribution of individual RNA sequences. Low entropy corresponds to better defined foldings. I will outline two possibilities.

  1. The biological reality is that there is essentially a unique structure. The fact that the computer predicts two or more structures is not a problem. They are close in free energy and are equally good solutions to the posed problem. Extra information is required to discover which structure is correct, or closer to correct.
  2. In reality, there are two or more structures. The software always assumes a distribution of different foldings. It is not at all surprising that dominant structures change as a function of temperature. I have seen this many times. It might have biological significance as in the example above.

I would not worry about simulating RNA folding in the range of 15 to 50° C. Folding at very high temperature is a huge problem, because the predictions are that no folding will occur at (for example) 70° C. When doing this, it is important to concentrate on what structural changes occur as temperature rises. I cannot make any further comments unless I process your sequence myself. This looks like a very interesting case that might be a good test of new tools that I am developing.