Research

 

Autocatalysis, the basis of self-reproduction

Autocatalysis underlies self-reproduction in chemical and biological systems. Autocatalysis is known to occur in various chemistries (small molecules, biopolymers) but we still lack a general understanding of it. We have recently shown the existence of stoichiometric motifs universal to all autocatalytic reaction networks. From this basis, we aim to systematically characterize autocatalysis in chemistry, thus uncovering the diversity of natural metabolisms and design artificial ones.

See our article in PNAS 2020.


Sequence-structure-activity relationship in self-reproducing RNAs

How probable is the appearance of self-reproduction during the origin of life? RNAs can self-reproduce by catalyzing the assembly RNA fragments into copies of themselves. However, only one such RNA sequence is known and has been engineered in the lab. To know whether self-reproduction is a widespread property in the space of RNA sequences, we leverage the diversity of catalytic RNAs (group I introns) found in the tree of life and experimental evolution combined with artificial intelligence. Back-and-forth iterations between generative models and experiments allow us to establish a picture of the neutral space of self-reproduction (meaning all sequences with this capability).

This work is notably supported by the HFSP young investigator program, in collaboration with Matteo Smerlak lab (Max Planck Institute for Mathematics in the Sciences, Leipzig, Germany), Eric Hayden lab (Boise State University, USA), Arati Ramesh lab (National Center for the Biological Sciences, Bangalore, India), and Martin Weigt lab (LCQB, Sorbonne Université).


The emergence of evolution in the RNA world

Evolution is the major driving force of biological complexity, but evolution is in itself a complex process that does not seem likely to appear in a spontaneous manner. Thus, how evolution can emerge from purely physical-chemical processes remains a fundamental puzzle. We study this question in in vitro RNA reaction networks capable of self-reproduction. The first step is to understand how the Darwinian properties of variation, reproduction with heredity, and selection can be implemented from molecular mechanisms. We notably investigate the role of recombination as a driving molecular mechanism. Evolution further requires its different properties to work in concert. This must rely on a form of a compartmentalized life cycle. Indeed, each compartment provides a unit of selection for collectives of molecules and natural selection operates on populations of such compartments across generations. To create and analyze compartments, we make use of droplet microfluidic technologies (in collaboration with Andrew Griffiths, Laboratory of Biochemistry ESPCI) and self-assembled compartments (in collaboration with Tommaso Fraccia, SAM team).

See our articles in Science 2016, Physical Review Letters 2018, and Nature Communications 2021. For conceptualization of the origin of life, see iScience 2020.

This work is supported by the ERC consolidator AbioEvo 2021-2026.


Evolutionary constraints in gene networks

Cellular activity relies on thousands of genes organized in networks of functional interactions. On one hand, gene networks are shaped by evolution. On the other hand evolution is constrained by the structure of gene networks. We aim to uncover the interplay between gene networks and evolution from the mechanistic level. For instance, given genes that interact in a metabolic pathway or regulate each other, do certain mutations alter the effect of each other on a phenotype (a phenomenon referred to as epistasis)? Until recently, we have focused on detailed experimental studies on well-characterized gene networks of small size, interpreted in the light of theoretical models of biochemical interactions. We are currently expanding the scope and generality of our approach at the cellular scale. For this, we develop strategies to measure rich phenotypes (e.g. transcriptomes, fitness) in a large number of variants of microorganisms using microfluidics and sequencing. We also study how mechanistic models can be combined with machine learning or data resulting from high-throughput screens.

See our articles in Nature Communications 2018 and Science Advances 2020

Collaboration with Sander Tans (AMOLF), and Olivier Tenaillon (INSERM, Hôpital Bichat).


Response and adaptation to complex environments

Integrating multiple signals from the environment is critical for the survival of organisms. Certain signals may elicit a specific response that help organisms to benefit from an environmental change (e.g. expression of a metabolic operon in the presence of a certain carbon source). Others may be harmful to the cell and cause a generic stress response or merely lead to death (e.g. antibiotics). We aim to understand how environmental signals perturb gene expression and in turn, how genetic evolution reshapes responses to the environment. We are currently focusing on antibiotic interactions, where certain combinations have a stronger effect than expected from their effect alone (called synergistic interactions), a weaker effect (antagonistic interactions). For this, we are developing experimental strategies to impose a large number of drug combinations of drugs in microfluidic devices, and measure their effect on cell growth.

See our articles in PNAS 2020, Cell Systems 2020


Development and application of millifluidic technologies to the study of microbial population biology

Working with LCMD and MilliDrop Instruments, LGE is at the forefront of the development of new applications for droplet technologies in microbiology and evolution. Initial foci of investigation include the relationship between microbial population density and growth dynamics, interactions between populations of pyoverdin-producing and non-producing populations, and phage-bacteria co-evolution.


Ecological scaffolding and the major evolutionary transitions

Life is hierarchically structured, with replicating entities nested within higher order self-replicating structures. Take, for example, multicellular life: the multicellular entity replicates, as do the cells that comprise the organism. Inside cells are mitochondria that also have capacity for autonomous replication; the same is true of chromosomes within the nucleus, and of genes that comprise chromosomes. Such hierarchical structure reflects a series of major evolutionary transitions in which lower order self-replicating entities have been subsumed within higher order structures. Typically this involves the lower level entity “giving up its right to autonomous replication” and with this “sacrifice” comes enslavement to the “needs” of the higher order “corporate body”. Posed in these terms it is difficult to see how evolutionary transitions unfold; how selection might shift levels and why life is hierarchically structured.

Necessary for progress is clarity concerning what needs to be explained: the evolution of Darwinian Individuality those properties of entities (variation, reproduction and heredity) that ensure participation in the process of evolution by natural selection. There has been a tendency to assume these properties as pre-existing, but they are not: they are derived and require evolutionary explanation. Pressing to the heart of the problem, the challenge is to explain how Darwinian properties emerge from non-Darwinian entities by non-Darwinian means. This challenge permeates each evolutionary transition including the emergence of life from matter.

Solutions to this seemingly unsolvable problem arise once ecology is considered and thus possibilities for Darwinian properties to be exogenously imposed. Such ideas underpin on-going work on the evolutionary transition to multicellularity and also the construction of symbiotic communities. A major motivation for development of millifluidic technologies stems from the realisation that droplets themselves can become units of selection in their own right. This has driven theoretical developments that guide operation of selection over longer time scales, investigations into the emergence of heredity, reproduction and new opportunities for top-down engineering of microbial communities, with applications in biotechnology, medicine and agriculture.

See publication in Nature Ecology & Evolution (2020) here.


The evolution of heredity

In major evolutionary transitions of the egalitarian type, the primary challenge is to explain how alignment of reproductive interests of the ancestral partners came into being given destabilising effects wrought by individual-level selection. One possible explanation involves growth of partner types under ecological conditions (specific population structures) that result in collectives of types being units of selection.

Our focus is on development of mathematical descriptions of evolution within this meta population structure and subsequent testing of ideas using millifluidic devices.

Guilhem Doulcier worked on this in collaboration with Silvia De Monte (EEM, IBENS), and Amaury Lambert (SMILE, UPMC). The work was funded by the Programme Origines et Conditions d’Apparition de la Vie PSL Research University, Paris. (ANR-10-IDEX-0001-02)

See publication in eLife (2020) here.


Lateral gene transfer and impacts on the ecological and evolutionary dynamics of microbial communities

Although the mechanics of horizontal gene transfer are well understood, its operation, impact and dynamic at the level of communities is unknown. In a year-long selection experiment we have followed "adaptation" of microbial communities to a new carbon source in the presence and absence of horizontal gene transfer. Current efforts focus on dissecting the evolutionary response using metagenomic approaches incorporating chromosome conformational capture and whole genome resequencing, and stable isotope analysis.

See publications in Philosophical Transactions of the Royal Society (2020) here and here.


Biophysics of microbial mat formation

At the heart of many Rainey lab experimental analyses of evolution in real time are a class of genotype that builds a complex self-supporting mat at the air-liquid interface of static broth microcosms. New quantitative optical scanning approaches have been developed by LGE to study the dynamics of mat formation, and to unravel hitherto unstudied dimensions of the genotype-to-phenotype map.

Maxime Ardré is working on this.

See publication in Journal of Bacteriology (2019) here.


Phage in droplets

Phage drive the evolution of bacteria through antagonistic coevolution and by promoting transfer of genetic material between bacteria. This project seeks to understand the mechanism behind a phage-induced growth phenotype using and Evolution Machine, developed by Millidrop Instruments.


The biophysics of diffusible products

Bacteria produce and secrete a range of molecules many of which affect interactions with neighbouring cells. A particularly well-studied example is the water soluble iron-chelating agent pyoverdin. Development of new millifluidic technologies in combination with microscopy and genetics allows enhanced understanding — and quantification of — the physical basis of microbial interactions.

Clara Moreno-Fenoll and Maxime Ardré are working on this with Clara focused on the mechanistic basis of pyoverdin privatisation


Building hybrid microbe-machine ecosystems

The interface between artificial intelligence and factors governing the establishment of biological complexity is largely unexplored. LGE is home to a new programme in which machines manipulate microbial systems, monitor their responses at microscopic scale, processes the resulting response, and then further impose new stimuli to control pattern formation. Ultimately we seek an artificial intelligence that will co-evolve with a given microbial community.


For the projects of the MMN and SAM teams, please visit:

Top