Modelling Translation Initiation under the Influence of sRNA

Fabian Amman, Christoph Flamm, Ivo L. HofackerInternational Journal of Molecular Sciences 2012 · DOI

TL;DR

This paper introduces a novel computational model that simulates how bacterial small non-coding RNAs (sRNAs) affect translation initiation. By integrating information on ribosome-mRNA interactions, sRNA-mRNA binding, and gene expression data, the model can predict whether an sRNA will activate or repress translation of its target mRNA, significantly improving the accuracy of sRNA target prediction and reducing the need for costly experimental validation.

What Problem Does This Paper Address?

Bacterial small non-coding RNAs (sRNAs) are crucial regulators of gene expression, acting at the post-transcriptional level. While the number of identified sRNAs has grown rapidly, their functional characterization and the identification of their target mRNAs remain a significant bottleneck. Existing computational methods for predicting sRNA-mRNA interactions often suffer from a high rate of false positives, making it difficult to prioritize potential targets for experimental validation. This paper aims to bridge this gap by developing a model that not only predicts sRNA-mRNA binding but also evaluates the functional consequence of this binding on translation initiation, a key step in gene expression.

Methods

The core of this work is a qualitative model for simulating translation initiation. This model builds upon a previously established mathematical framework for ribosome-mRNA interaction and expands it to incorporate the influence of sRNA binding. The key components of the model are:

  • Ribosome Recognition Site (RRS) and Ribosome Docking Site (RDS): The model first identifies the energetically most favorable binding site for the ribosome on the mRNA (RRS) and defines the region of the mRNA that the ribosome occupies during initiation (RDS). This involves calculating hybridization energies between the 16S rRNA and the mRNA.
  • sRNA-mRNA Interaction: The model predicts potential sRNA binding sites on the mRNA and calculates the binding energy.
  • Accessibility: Crucially, the model considers the accessibility of the RDS. It accounts for how sRNA binding can either compete with ribosome binding or alter mRNA structure, thereby affecting RDS accessibility. This is achieved using the RNAup program to calculate probabilities of the RDS being exposed.
  • Kinetic Modeling: The model translates these interactions into a system of chemical reaction equations that describe the equilibrium concentrations of various species (free mRNA, sRNA-bound mRNA, ribosome-bound mRNA, and sRNA-ribosome-bound mRNA).
  • Classification: The model calculates the ratio of translationally active mRNA in the presence and absence of the sRNA. A fold change greater than 2 (absolute value) is used as a threshold to classify a potential interaction as functionally significant.

The implementation relies on several modules from the ViennaRNA Package:

  • RNAplex: For finding potential sRNA-mRNA interaction sites.
  • RNAduplex: For calculating ribosome-mRNA interaction energies and identifying RRS/RDS.
  • RNAup: For determining RNA accessibility probabilities.
  • RNAplfold: Used in conjunction with RNAplex to consider accessibility effects.

Datasets

The study utilizes publicly available deep sequencing data for Escherichia coli (specifically, E. coli MG1655 grown in rich media) from the Sequence Read Archive (submission ID: SRA050648). This data was used to estimate mRNA expression levels and, consequently, the concentrations of reactants in the model. The study also references the EcoCyc database for information on E. coli trans-acting ncRNAs and their targets.

Key Insights and Ideas

  • Beyond Binding: Functional Consequence Matters: The most significant insight is the shift from simply predicting sRNA-mRNA binding to modeling the functional outcome of that binding on translation initiation. This moves beyond static interaction predictions to a dynamic, mechanistic understanding.
  • Accessibility is Key: The model highlights the critical role of mRNA accessibility, particularly the Ribosome Docking Site (RDS), in translation initiation. sRNAs can exert their influence by either directly blocking this site or by inducing structural changes that affect its accessibility.
  • Quantitative Prediction of Regulation Type: The model can distinguish between activation and repression of translation, providing a more nuanced understanding of sRNA function than simple binding prediction.
  • Significant Reduction in False Positives: By incorporating functional simulation, the model dramatically reduces the number of predicted targets, making experimental validation more efficient and cost-effective.

Key Results

  • The model successfully simulated thirteen out of fifteen tested known sRNA-mRNA interactions in a qualitative manner.
  • When applied to E. coli sRNA RyhB, the model reduced the number of potential targets from 1921 (based on simple binding energy) to 446, a reduction to less than one-fourth.
  • For RpoS translation regulation, the model identified six potential sRNA regulators, including three known ones and three novel candidates.
  • The study demonstrated that the energetically most favorable binding site is not always the one with the strongest functional impact on translation initiation.

Limitations and Future Work

  • Equilibrium vs. Kinetics: The current model is based on equilibrium states. The authors acknowledge that actual translation initiation is a dynamic, stochastic process that can occur co-transcriptionally, and kinetic effects might be more important for certain interactions.
  • Single Binding Site Assumption: The model currently considers only one sRNA binding site at a time and does not account for competition between multiple mRNA molecules for a limited pool of sRNA.
  • RNA Chaperone (Hfq) Ignored: The role of RNA chaperones like Hfq, which can influence sRNA-mRNA interactions, is not explicitly modeled. While Hfq might stabilize interactions, its direct impact on the thermodynamic equilibrium state is considered sufficient for detecting Hfq-dependent targets. However, a more comprehensive model could incorporate Hfq's effects on kinetics.
  • SRNA Concentration Estimation: The model relies on estimated sRNA concentrations, which can be a challenge due to limited experimental data. Future work could benefit from more comprehensive and condition-specific sRNA expression data.
  • Future Work: The authors plan to make their approach available as a web-based service integrated into the RNApredator tool for post-processing of sRNA target predictions. They also suggest that future work could explore the inclusion of Hfq's effects and a more detailed kinetic modeling approach.

Summary updated

Abstract

Bacterial small non-coding RNA (sRNA) plays an important role in post-transcriptional gene regulation. Although the number of annotated sRNA is steadily increasing, their functional characterization is still lagging behind. Various computational strategies for finding sRNA−mRNA interactions, and thus putative sRNA targets, were developed. Most of them suffer from a high false positive rate. Here, we present a qualitative model to simulate the effect of an sRNA on the translation initiation of a potential target. Information about the ribosome−mRNA interaction, sRNA−mRNA interaction and expression information from deep sequencing experiments is integrated to calculate the change in translation initiation complex formation, as a proxy for translational activity. This model can be used to post-evaluate predicted targets, hence condensing the list of potential targets. We show that our translation initiation model, under the influence of an sRNA, can successfully simulate thirteen out of fifteen tested sRNA−mRNA interactions in a qualitative manner. To show the gain in specificity, we applied our method to a target search for the Escherichia coli sRNA RyhB. Compared with simple target prediction without post-evaluation, we reduce the number of targets to less than one fourth potential targets, considerably reducing the burden of experimental validation.
BibTeX
@article{amman2012modelling,
  author = {Fabian Amman and Christoph Flamm and Ivo L. Hofacker},
  title = {{Modelling Translation Initiation under the Influence of sRNA}},
  journal = {International Journal of Molecular Sciences},
  year = {2012},
  doi = {10.3390/ijms131216223},
}
Figure