RNA Kinetics: Theory

In our model for RNA secondary structure formation we assume that each of the candidate helices (see definition in the main text) may exist in two states. A helix can be completely zipped, or it can be completely decayed. This effectively targets the time scale of RNA secondary structure transitions that is of order of 10^-3sec. The following schema is used to estimate the effective transition constants between zipped and decayed states:

· Spontaneous decay of a helix occurs as a set of consecutive unzipings of the terminal nucleotide pairs. It can be thought of as a branching Markov process (Figure 1).

· Formation of a new helix occurs in two steps: formation of the loop and zipping of the helix. To determine the kinetic constants, we use the transition state theory, which is often used in chemical kinetics (Figure 2).

Figure 2. Transition states of a helix formation process.

where

is probability of i-th transition state and k

is an elementary event constant. Suppose that all the transition states probabilities are equal. Thus,

here

denotes the stacking pairs number in the helix. The probability to form a loop relates to the entropy loss as

One can obtain the same formulae considering helix decay as a branching Markov process (Figure 1). Thus the physical sense of the constant

can be clarified. Thise constant is the constant of the last nucleotide pair locking on the helix growths. Its value is estimated as 1.e+6 .. 1.e+8 (Porske, 1974). Here is a table of characteristic times for some events given .

Transition	time
A formation of a 5-nucleotide helix with a hairpin loop of 6 nucleotides (i.e. the most favorable helix)	2.e-4
A spontaneous decay of a helix containing 5 A-U nucleotide pairs	8.e-4
A spontaneous decay of a helix containing 5 G-C nucleotide pairs	1.4e+4
A spontaneous decay of a helix containing 8 G-C nucleotide pairs	4.7e+10=1500 years!!!

Basic model for kinetics of RNA folding .

The RNA secondary structure formation can be represented as a Markov process. The states of the process are the secondary structures, the transitions are a new helix formation or an existing helix decay. As far as the transitions are uniform (i.e. at any time, their probabilities do not depend on the time count start), ordinary (i.e. only one transition is possible at any time), and history-independent (transitions probabilities depend only on current state rather than on pre-history), the transition time is a random, exponentially-distributed random value

The modeling process is the following.
1. We enumerate all the potential (candidate) helices with lifetime that exceed a given threshold.
2. We initiate the process with a trivial structure with no helices.
3. For the current state (structure), we form a complete list of potential transitions with their probabilities from all existing helices decay, all possible helices (i.e. that do not contradict with any of existing helices) formation, and the chain elongation.

4. The kinetic constants are calculated for these transitions.
5. We choose the transition to execute. A random value drawn uniformly from the interval wher

and then the number of the transition is defined from

The time between the transition and the pervious one is drawn from the exponential distribution

5. The chosen transition is executed and the experiment time is incremented by .
6. If the current time is still less than a pre-given value of the overall experiment time, return to 3.
7. If the times of runs we made is less than a pre-given number N, return to 2.
The result of the algorithm work is kinetic ensemble of secondary RNA structures, i.e. a set of secondary structures with their time-dependent probabilities. The relative precision of the probabilities evaluation can be estimated by as . To include the RNA molecule synthesis, it is enough to start from zero length chain on step 2 and to add the chain elongation by 1 nucleotide as a possible transition. The constant for this transition is usually .

The refined model

Virtually, the helices are not independent. They can overlap. Let's consider a classification of mutual positions of two candidate helices (Figure 4).

· Complete compatibility (A). The second helix can form without the first helix decay.
· Partial compatibility (B). The helix 2 can be freely initiated, but the there are helix positions overlap, and a sliding loop is formed between two helices.
· One of the strands of helix 2 is located inside a strand of helix 1 (C). It is necessary to untwist helix 1 partially to initiate helix 2. After the initiation, the helix 2 will be supplanted by helix 1. If helix 2 exists, helix 1 can be freely initiated and then it will supplant helix 2.
· It is necessary to untwist helix 1 partially to initiate helix 2 (D). After it, 1 will supplant 2 and nothing happens.
· The helix 2 can be initiated freely (E). Then, helix 2 supplants helix 1 and there is no helix 1 in the final structure.
· A pseudo-knot (F). In this case, to initiate helix 2 it is necessary either to overcome the pseudo-knot formation energetic threshold or to decay helix 1 completely.
Another characteristic of the refined model is that we do not treat helix decay as a sufficient event, because a new helix can form (possibly, the same) after decay. The process of new helix formation given that it is hampered by other helices can be represented as a set of elementary events.

There are several ways to form the new helix, though the final structure is the same for all of them. All the events except the last increase the free energy, so we can evaluate the transition constant as

(3)

here G is the transition free energy, is constant of a reverse from the final state to the to one of the pre-final (threshold) states, i.

Let's verify the formula. A threshold transition occurs, thus the flux is summarized over all fluxes from all the pre-final states,

where

is probability of i-th pre-final state,

is the constant of the transition from the state to the final one. The flux is relatively small, so we can suppose that is close to the balanced value

where

is energy of i-th pre-final state relative to the final state. The Van't Hoff's law requires:

whre in

energy of -th pre-final state relative to the final state. Thus, we obtain the formula (3).

Reference

Porshke D. (1974). A direct measurement of the unzippering rate of a nucleic acid double helix. Biophys.Chem.. 2: 97-101

decay of all hampering helices	the free energy increases
new helix initiation	the free energy increases
the helix formation completing	the free energy decreases

RNA Kinetics theory

Contents:

Kinetic constants estimation

Basic model for kinetics of RNA folding

The refined model

Estimation of kinetic constants

Figure 1. A helix decay process.

Figure 2. Transition states of a helix formation process.

(1)

(2)

Basic model for kinetics of RNA folding .

Figure 3 The definition of possible transitions.

The refined model

Figure 4 Possible mutual positions of two candidate helices.

(3)

Reference