Elsevier

Biosystems

Volume 88, Issues 1–2, March 2007, Pages 163-172
Biosystems

Alternative routes and mutational robustness in complex regulatory networks

https://doi.org/10.1016/j.biosystems.2006.06.002Get rights and content

Abstract

Alternative pathways through a gene regulation network connect a regulatory molecule to its (indirect) regulatory target via different intermediate regulators. We here show for two large transcriptional regulation networks, and for 15 different signal transduction networks, that multiple alternative pathways between regulator and target pairs are the rule rather than the exception. We find that in the yeast transcriptional regulation network intermediate regulators that are part of many alternative pathways between a regulator and target pair evolve at faster rates. This variation is not solely explicable by higher expression levels of such regulators, nor is it solely explicable by their variable usage in different physiological or environmental conditions, as indicated by their variable expression. This suggests that such pathways can continue to function despite amino acid changes that may impair one intermediate regulator. Our results underscore the importance of systems biology approaches to understand functional and evolutionary constraints on genes and proteins.

Introduction

Genetic changes in the smallest parts of molecular networks – genes and proteins – can affect the structure of these networks. Conversely, this structure may itself constrain properties of genes and proteins, and the kinds of mutations they can tolerate. To ask how biological networks constrain their parts has become possible with the availability of experimental data on genome-scale metabolic, transcriptional regulation, and protein interaction networks (Forster et al., 2003, Ito et al., 2001, Lee et al., 2002, Uetz et al., 2000, von Mering et al., 2002). If network structure constrains network parts, then an understanding of the evolution of genes and proteins will require an understanding of molecular networks. We here focus on one aspect of network organization, alternative pathways through a genetic network, and how such pathways affect genes in a transcriptional regulation network. The importance of alternative pathways is hinted at by systematic studies on metabolic networks, where alternative pathways of metabolite flow can make a network robust against loss-of-function mutations in enzymes (Edwards and Palsson, 2000, Segre et al., 2002). However, no comparable information exists for any regulatory network.

The evidence for alternative or ‘redundant’ pathways through regulatory gene networks is mostly anecdotal or circumstantial (Bi et al., 2000, Ho and Satoh, 2003, Kolodner et al., 2002, Lefers et al., 2001, LeRoith, 2000, Morris et al., 1995, Passalaris et al., 1999, Vance and Wilson, 2002, Wang et al., 2002a). This evidence typically comes from molecular biological studies where different regulatory pathways, sometimes involving overlapping sets of regulators, can influence the same genes or biological processes. For example, the RAD17 gene product of the yeast Saccharomyces cerevisiae is involved in detecting DNA damage during different stages of the yeast cell cycle. This protein interacts with different sets of other regulatory proteins, which are part of several redundant regulatory pathways that ensure genome stability (Kolodner et al., 2002). Another such example comes from the degradation of cholesterol in mice. Bile acids, the degradation products of cholesterol, indirectly repress the genes necessary for cholesterol degradation. They do so through several incompletely characterized alternative pathways that involve different transcription factors (Wang et al., 2002a).

Available data on genome-scale biological networks elucidated by currently available functional genomic techniques is by necessity incomplete. It may incorporate only one mode of regulation, such as transcriptional regulation; it may not contain information about all relevant regulatory molecules; and it may contain substantial experimental noise. Even with these caveats, however, such data allows a more systematic exploration of important systems – biological questions than small-scale data on individual pathways. Is the existence of alternative pathways between a regulator and its (indirect) regulatory target the exception or the rule? What are the consequences of such alternative pathways for the function and evolution of intermediate regulators, the regulatory molecules that stand between the regulator and its target? These are some of the questions we will address here.

Section snippets

Alternative regulatory pathways are not rare

We find that in regulatory networks of even moderate complexity, many pairs of regulatory molecules and their targets are connected by more than one regulatory pathway. Fig. 1a illustrates this notion with data from the transcriptional regulation network of the yeast S. cerevisiae (Lee et al., 2002). In this network, a directed edge (link) connects two genes A and B if A encodes a transcriptional regulator that can regulate the expression of B, as indicated by its binding to the regulatory

Conclusions

We show that alternative pathways are abundant in more than a dozen biochemical regulatory networks. For the transcriptional regulation network of S. cerevisiae, molecular evolution data suggest that such alternative pathways may provide robustness to mutation. Mutational robustness in a network may either be an adaptation in and by itself, or it may have emerged as a by-product of other evolutionary processes. To distinguish between these possibilities remains an important task for future

Signal transduction networks

The science signal transduction knowledge environment (http://stke.sciencemag.org/cgi/cm) contains a collection of signal transduction pathways manually assembled by experts on these networks. We analyzed the structure of all 15 signal transduction networks with more than 30 nodes that were available in this repository in May 2004. These networks are the adrenergic pathway (http://www.stke.org/cgi/cm/; CMP_8762), a network that mediates the responses of cells to epinephrine and norepinephrine;

Acknowledgments

AW would like to acknowledge support through NIH grant GM63882 to the University of New Mexico, the continued support of the Santa Fe Institute, as well as the support of the Institut des Hautes Etudes Scientifique for a sabbatical stay.

References (42)

  • M. Lefers et al.

    Genetic dissection of the drosophila Cubitus interruptus signaling complex

    Dev. Biol.

    (2001)
  • L. Wang et al.

    Redundant pathways for negative feedback regulation of bile acid production

    Dev. Cell

    (2002)
  • E. Bi et al.

    Identification of novel, evolutionarily conserved Cdc42p-interacting proteins and of redundant pathways linking Cdc24p and Cdc42p to actin polarization in yeast

    Mol. Biol. Cell

    (2000)
  • W. Blake et al.

    Noise in eukaryotic gene expression

    Nature

    (2003)
  • S. Chu et al.

    The transcriptional program of sporulation in budding yeast

    Science

    (1998)
  • J.L. DeRisi et al.

    Exploring the metabolic and genetic control of gene expression on a genomic scale

    Science

    (1997)
  • R.B. Dunn et al.

    Selection for an invariant character – ‘vibrissa number’ – in the house mouse

    Nature

    (1958)
  • J.S. Edwards et al.

    The Escherichia coli MG1655 in silico metabolic genotype: its definition, characteristics, and capabilities

    Proc. Natl. Acad. Sci. USA

    (2000)
  • M.B. Eisen et al.

    Cluster analysis and display of genome-wide expression patterns

    Proc. Natl. Acad. Sci. USA

    (1998)
  • M. Elowitz et al.

    Stochastic gene expression in a single cell

    Science

    (2002)
  • A. Evangelisti et al.

    Molecular evolution in the transcriptional regulation network of yeast

    J. Exp. Zool./Mol. Dev. Evol.

    (2004)
  • J. Forster et al.

    Genome-scale reconstruction of the Saccharomyces cerevisiae metabolic network

    Genome Res.

    (2003)
  • H.B. Fraser et al.

    A simple dependence between protein evolution rate and the number of protein–protein interactions

    BMC Evol. Biol.

    (2003)
  • A.P. Gasch et al.

    Genomic expression programs in the response of yeast cells to environmental change

    Mol. Biol. Cell

    (2000)
  • S. Ghaemmaghami et al.

    Global analysis of protein expression in yeast

    Nature

    (2003)
  • M. Hahn et al.

    Molecular evolution in large genetic networks: does connectivity equal importance?

    J. Mol. Evol.

    (2004)
  • E. Ho et al.

    Repair of single-strand DNA interruptions by redundant pathways and its implication in cellular sensitivity to DNA-damaging agents

    Nucleic Acids Res.

    (2003)
  • T. Ito et al.

    A comprehensive two-hybrid analysis to explore the yeast protein interactome

    Proc. Natl. Acad. Sci. USA

    (2001)
  • I.K. Jordan et al.

    Correction: no simple dependence between protein evolution rate and the number of protein-protein interactions: only the most prolific interactors evolve slowly

    BMC Evol. Biol.

    (2003)
  • M. Kellis et al.

    Sequencing and comparison of yeast species to identify genes and regulatory elements

    Nature

    (2003)
  • M. Knop et al.

    Epitope tagging of yeast genes using a PCR-based strategy: more tags and improved practical routines

    Yeast

    (1999)
  • Cited by (37)

    • The Importance of Genetic Redundancy in Evolution

      2020, Trends in Ecology and Evolution
      Citation Excerpt :

      Origin of redundancy through duplication has received special attention, given that the genome-wide rates of gene duplication in multicellular eukaryotes have been observed in the range of 10–7 duplications per gene per generation [42] and locus-specific rates have been observed to be as high as 10–5 duplications per gene per generation [43,44]. For more comprehensive discussions of what we term ‘functional redundancy’ and related concepts, please see [9,33,41,45–47]. Genotypic redundancy refers to the total number of possible causal genotypes that could occur through mutation that yield a given phenotype (scaled by likelihood of occurrence, see later), while segregating redundancy refers to the realized number of different genotypes with the same phenotype that are present at a particular time point in the population or species.

    • The origins of mutational robustness

      2015, Trends in Genetics
      Citation Excerpt :

      The number of different phenotypes emerging from new mutations will increase with the number of neutral genotypes (i.e., possible backgrounds) in the population. Consequently, if there are a large number of neutral genotypes, the population can access more new phenotypes [2,17,18]. Thus, the larger the genotypic network (i.e., the larger the mutational robustness of a population), the higher the potential is to produce novel adaptations [19].

    • Systems-Level Response to Point Mutations in a Core Metabolic Enzyme Modulates Genotype-Phenotype Relationship

      2015, Cell Reports
      Citation Excerpt :

      In contrast to earlier studies, the focus of the present work is on the systems-level proteome and transcriptome response to the minimal and most fundamental genetic perturbations—missense point mutations introduced through genome editing into a core metabolic enzyme. A popular conceptual view in systems biology postulates that modularity and stability of transcriptional networks had evolved to confer robustness to biological systems (Bornholdt and Sneppen, 2000; Wagner and Wright, 2007). In particular, an effect of a point mutation in a robust biological system should be limited to genes and their protein products that physically, genetically, or metabolically interact with a mutated protein.

    • The Evolution of Intelligent Developmental Systems

      2013, Advances in Child Development and Behavior
      Citation Excerpt :

      The pathways are interlinked in elaborate networks, so that the reactions in one are contextualized by the reactions in many others (Peisajovich, Garbarino, Wei, & Lim, 2010). In this, cross-talk enzymes and other cell components have activities conditioned by others at potentially many statistical depths to reflect the deeper structure in experience (Wagner & Wright, 2007). Such structural assimilation has been assisted by a proliferation, in the course of evolution, of signaling proteins with increased numbers of “interaction domains,” facilitating recombination in an increasing variety of ways (Peisajovich et al., 2010).

    View all citing articles on Scopus
    View full text