MicroRNA degeneracy and pluripotentiality within a Lavallière-tie architecture confers robustness to gene expression networks

Abstract

Modularity, feedback control, functional redundancy and bowtie architecture have been proposed as key factors that confer robustness to complex biological systems. MicroRNAs (miRNAs) are highly conserved but functionally dispensable. These antinomic properties suggest that miRNAs fine-tune gene expression rather than act as genetic switches. We synthesize published and unpublished data and hypothesize that miRNA pluripotentiality acts to buffer gene expression, while miRNA degeneracy tunes the expression of targets, thus providing robustness to gene expression networks. Furthermore, we propose a Lavallière-tie architecture by integrating signal transduction, miRNAs and protein expression data to model complex gene expression networks.

Fig. 1

Bowtie architecture of metabolic networks. This is a modified Figure from Csete and Doyle [6]. Bowtie architecture is a combination of two modules coupled via a central element or core, which is defined as a reduced set of building blocks. All modern technologies, from manufacturing to the Internet, are organized with bow ties architecture. On the left hand side of the structure (pink wing), nutrients are catabolized into a few precursors (sugars, amino acids, nucleotides, fatty acids, and co-factors) that will be used as energy and building blocks for the cellular anabolism on the right hand side of the structure side (blue wing). The whole process is tightly regulated and controlled through different systems (black and blue arrows). In the “anabolism” wing (blue wing) the gene expression network structure is represented. It also exhibits bowtie architecture, organized around a core the Trans* (transcription/translation) machinery composed of few polymerases and universal components enabling gene transcription and translation

Fig. 2

The miRNA network. This graph represents the underlying organization, of the miRNA network we have recently characterized [1]. The community of miRNAs in red or module 2 (lower part of the network), targets primarily protein-coding genes involved in signal transduction, particularly small GTPase signaling, would fan in information from the cell microenvironment. The upper part of the network, the blue community or module 1, regroups miRNAs that primarily target transcriptional regulators, fanning out gene expression toward diverse phenotypic outcomes in response to input signals. Central to the network, several miRNAs (olive colored nodes) are connected to either community but do not target any particular function in the cell. The modularity of this network and its underlying organization resemble bowtie. However, the nature of the core (olive colored nodes) composed of miRNA and not of a reduced set of building blocks, does not fit the definition of bowtie architecture [6]

Fig. 3

Lavallière-tie structure to model the integration of the miRNA network into a larger gene expression network. The miRNA network is nested within a noncoding RNA (ncRNA) wing and regulate gene expression by acting both at the signal transduction level (red arrows) and at the Trans* (transcription/translation) machinery level (blue arrow) through feedback loops. The core module of the Lavalliere-tie is composed of few polymerases and universal components (nucleotides, aminoacids, ribosomes units) allowing transcription of both coding and non-coding genes and translation of protein coding genes. While three wings of the structure are composed of molecular objects, the fourth wing manages the flow of information coming from the microenvironment of the cell, through signal transduction and epigenetics mechanisms