To fully understand living systems we need (i) experimental techniques to describe them as accurately and comprehensively as possible, and (ii) computational models able to predict their evolution from a given state and in response to external signals.
The Imaging and Modeling Unit of Institut Pasteur, develops computational and experimental approaches to characterize and quantitatively predict selected cellular processes. Our current projects concentrate on : (i) investigating the dynamic spatial architecture of the genome and its functional consequences, (ii) developing high resolution or high throughput imaging techniques, and applying them to study genome architecture and the cell biology of pathogens, especially HIV, and (iii) leveraging the power of artificial intelligence (deep learning) to analyze biomedical imaging data. Our lab mobilizes a spectrum of expertise including biophysics, microscopy, informatics and cell biology, and works in close collaboration with several experimental groups, many of them at Institut Pasteur.
Nuclear architecture and function
The one-dimensional sequence of the genome is carried by long polymers (chromosomes), which fold in the three dimensional volume of the nucleus. How this folding occurs and what it implies for genome function remains largely unknown. To better describe genome architecture in yeast, we developed a method that maps the nuclear territories of fluorescently tagged chromatin loci by computationally analyzing images of thousands of cells (Berger et al. 2008). This revealed a strong territorial organization of the yeast nucleus (Figure 1, left). Application of this method to most yeast telomeres allowed us to identify some of the main determinants of locus positioning: genomic distance to centromere and hindrance by the nucleolar compartment (Thérizols et al. 2010).
To better understand the mechanisms of chromosome folding, we developed a simple predictive model of chromosome dynamics based on polymer physics (Figure 1, right). The model’s predictions are in very good quantitative agreement with our imaging data and with genome-wide DNA contact frequencies measured biochemically (Wong et al. 2012, 2013). This suggests that large-scale genome architecture in yeast is governed mainly by generic properties of randomly moving chromosomes rather than by DNA-specific factors. We also extended our modeling approach to infer key physical parameters of the chromatin fiber and to predict locus dynamics (Arbona et al. 2017). We now explore the consequences of this model for a quantitative understanding of functional processes, and in particular DNA repair (Agmon et al. 2012; Wong et al. 2013; Herbert et al. 2017; Fabre & Zimmer 2018).
Fluorescence microscopy is omnipresent in cellular biology but the resolution of standard microscopes is limited to ~200 nm, thus preventing detailed analyses of molecular structures. We are actively interested in single molecule localization microscopy (“pointillism”), which enables greatly improved resolution. We have implemented pointillist super-resolution microscopy hardware and reconstruction software (Figure 2) and are applying these techniques in various collaborations to study pathogens in their cellular hosts, with a particular interest for HIV (Lelek et al. 2012; Lelek et al. 2015).
In parallel, we aim to push the limitations of super-resolution methods. One important limitation is the use of invasive fluorescent tags, which can perturb biological functions. We demonstrated localization microscopy with FlAsH-tetracysteine tagging (FlAsH-PALM), which allows to obtain <30 nm resolution images of delicate microbial proteins such as the HIV integrase, without disrupting their function (Lelek et al. 2012). Recently, we developed ZOLA-3D, a combined optical and computational technique that allows to obtain 3D super-resolution images without scanning over up to 5 micrometers depth (Aristov et al. 2018). We also developed ANNA-PALM, a method based on deep learning that vastly accelerates image acquisition in localization microscopy and opens new doors to high throughput and live cell super-resolution microscopy (Ouyang et al. 2018).
Keywords: computational cell biology, super-resolution microscopy, deep learning, nuclear architecture, yeast, host-pathogen interactions, HIV.
Agmon N, B. Liefshitz, C. Zimmer, E. Fabre, and M. Kupiec, “Effect of nuclear architecture on the efficiency of double-strand break repair,” Nature Cell Biology, vol.15, no. 6, pp. 694-9 (2013).
Arbona JM, Herbert S, Fabre E, C Zimmer. Inferring the physical properties of yeast chromatin through Bayesian analysis of whole nucleus simulations. Genome Biology. 18:81, doi: 10.1186/s13059-017-1199-x (2017).
A. Aristov, B. Lelandais, E. Rensen, C. Zimmer. ZOLA allows flexible 3D localization microscopy over an adjustable axial range. Nature Communications, 36(5):460-468 (2018).
Berger A, G. G. Cabal, E. Fabre, T. Duong, H. Buc, U. Nehrbass, J. C. Olivo-Marin, O. Gadal, and C. Zimmer, “High-resolution statistical mapping reveals gene territories in live yeast,” Nature Methods, vol. 5, no. 12, pp. 1031–1037, 2008.
Fabre E & C Zimmer. From dynamic chromatin architecture to DNA damage repair and back Nucleus, Vol. 9, no. 1, 161–170 (2018).
Herbert S, A. Brion, J.-M. Arbona, M. Lelek, A. Veillet, B. Lelandais, J. Parmar, F. Fernandez, E. Alamyrac, Y. Khalil, E. Birgy, E. Fabre, and C. Zimmer. Chromatin stiffening underlies enhanced locus mobility after DNA damage in budding yeast EMBO Journal. Sep 1;36(17):2595-2608 (2017).
Lelek M, F. Di Nunzio, R. Henriques, P. Charneau, N. Arhel, and C. Zimmer, “Superresolution imaging of HIV in infected cells with FlAsH-PALM.,” PNAS, vol. 109, no. 22, pp. 8564–9, May 2012.
Lelek M, Casartelli N, Pellin D, Rizzi E, Souque P, Severgnini M, Di Serio C, Fricke T, Diaz-Griffero F, Zimmer C, Charneau P, Di Nunzio F. Chromatin organization at the nuclear pore favours HIV replication. Nature Communications, Mar 6; 6:6483. doi: 10.1038/ncomms7483 (2015).
Ouyang W., A. Aristov, M. Lelek, X. Hao, C. Zimmer. Deep learning massively accelerates super-resolution localization microscopy. Nature Biotechnology, 2018.
Thérizols P, T. Duong, B. Dujon, C. Zimmer, and E. Fabre, “Chromosome arm length and nuclear constraints determine the dynamic relationship of yeast subtelomeres,” PNAS, vol. 107, no. 5, p. 2025, 2010.
Wong H, H. Marie-Nelly, S. Herbert, P. Carrivain, H. Blanc, R. Koszul, E. Fabre, and C. Zimmer, “A Predictive Computational Model of the Dynamic 3D Interphase Yeast Nucleus.,” Current Biology : CB, vol. 22, no. 20, pp. 1881–90, Oct. 2012.
Wong H, Arbona J-M, Zimmer C. "How to build a yeast nucleus." Nucleus. Aug 22; 4(5):1-6 (2013).