One unsolved challenge in biology is to discover how all cell constituents, including its genes, proteins and organelles interact to control cell behavior in real time and in vivo. We address this challenge by imaging molecular events and single-cell dynamics directly in the living mouse embryo. Our approach is enabling us to discover entirely new mechanisms and unexpected cellular structures controlling how the early mammalian embryo forms.

The mouse embryo is a phenomenal system to reveal the dynamics controlling fundamental cell functions in vivo. This embryo was microinjected with RNA to label the chromatin (green) and cell membranes (red) throughout preimplantation development

Transcription factors (TFs) are essential regulators of cell fate, yet their DNA–binding dynamics in single cells and in vivo had remained unknown. Therefore, we started to develop methods to study the DNA search mechanism and DNA–binding properties of TFs in complex systems like embryos and tissues.

We first took advantage of photo-actiovatable GFP (paGFP) to selectively uncage a defined pool of TFs  in single cells of live embryos and characterize their nucleo-cytoplasmic movement in single cells of live embryos. Right image shows selective activation of the TF Oct4 in a live embryo (Plachta et al, 2011, Nat. Cell Biol).

We have also advanced these investigations to the level of TF–DNA interactions. To achieve this, we combined selective photo-activation with fluorescence correlation spectroscopy (paFCS). This method enables us to quantify the DNA–binding properties of TFs in single cells of live embryos and tissues.

(A) Using paFCS), we uncage TFs fused to paGFP with multiphoton lasers deep in live embryos. (B) Controlling the number of fluorescent TFs by photo-activation makes FCS feasible in vivo. (C) Comparing normal and mutant TFs reveals the mechanisms controlling DNA–binding. In this example we studied the dynamics of the TF Oct4, a key regulator of pluripotency and embryogenesis.

(Kaur et al, 2013, Nat. Comms;  Zhao et al, 2017, Nat. Protocols)

More recently, we discovered that pluripotency-associated TFs like Sox2 display different DNA–binding properties as early as the 4-cell stage of development, long before cells determine their fate.


We also found how TF–DNA interactions are regulated by histone methylation, and that heterogeneities in DNA binding predict cell fate (White et al, 2016, Cell). 

VIDEO. Cell PaperFlick (White et al, 2016, Cell)

We also advance the application of imaging technologies to reveal how dynamic changes in the organization of the cell cytoskeleton control cell shape and cell polarization in vivo. The microtubule cytoskeleton of most animal cells is organized by an organelle called the centrosome.


We recently imaged microtubules in the early mouse embryo, which lacks centrosomes. We found that that the cells of the embryo are connected by a stable microtubule bridge (shown in this image).


This bridge structure functions as a non-centrosomal  microtubule organizing center (MTOC), directing the growth of microtubules within the cell. Moreover, the microtubules emanating from this MTOC transport key proteins to the cell membrane, including E-cadherin, to control cell polarization during early development. This study reveals a new form of cytoskeleton organization in vivo (Zenker et al, 2017, Science).


Live imaging enables us to discover new sub-cellular structures, such as the microtubule bridge found in this Video (left). We can then zoom into the cell and do the molecular analysis to reveal how this structure functions in vivo. The Video shows microtubule plus-ends growing out of the bridge into the cell.

Using 3D electron microscopy (with Rob Parton, UQ Australia) revealed vesicles transported along the microtubules emanating from this bridge. Some of these are endosomes delivering E-cadherin to the membrane.


We have also discovered new functions for the actin cytoskeleton.  We found that the cells of the early embryo form long protrusions (or filopodia) containing E-cadherin and F-actin, which they use to draw their neighboring cells closer to achieve polarization and compaction. Although filopodia are critical for wound healing, signalling and cancer, current systems to study cell protrusions in live mammalian cells are restricted to culture conditions. Therefore, we now study the filopodia of the early embryo to establish how mammalian cells regulate these protrusions in vivo.

The Video above shows a cell retracting its filopodia in a live embryo (Fierro-Gonzalez et al, 2013, Nature Cell Biology).

3. How cells choose their position:

Revealing the forces driving mammalian morphogenesis

Investigating how cells adopt their specific positions within a tissue or organ is key to understand how our body forms. We combine live imaging with computer segmentation methods to reveal how cells segregate to form the pluripotent inner mass in the living mouse embryo.

This inner mass acts as precursor for most cells in our body and is one of the first structures to form during development.

Graphical Abstract.

Samarage et al, 2015, Developmental Cell

Unlike models based on highly orientated cell divisions, we showed that the pluripotent mass is formed primarily by a morphogenetic process known as apical constriction.

Using advanced laser ablation techniques, we revealed how anisotropies in cortical tension –a force generated by contractility of the actomyosin cell cortex– drives this morphogenetic process (Samarage et al, 2015, Developmental Cell).

The images (left) show the main tensile forces driving cells inside the embryo to form the inner mass.

VIDEO.  Seeing how the pluripotent inner cell mass forms​ in real time.

Transformation from morula to blastocyst is a defining event of mouse and human development. During this transition, the embryos must establish a permeability barrier to enable expansion of the blastocyst cavity, yet the mechanism triggering this sealing had remained unknown.

F-actin rings in the early morula.

Preceding blastocyst stage, an F-actin ring forms at the apical pole of the cells of the embryo. Unlike stereotypical actin rings that constrict, our in vivo imaging reveals that these rings expand all the way to the cell-cell junctions where they couple and trigger a zippering process that seals the embryo to enable blastocyst formation (Zenker et al, 2018, Cell).

VIDEO.  Actin rings expand over the entire cell cortex  (Zenker et al, 2018, Cell).

VIDEOThe zippering process recruits the key components of adherens and tight junctions. The video shows formation of tight junctions (labeled by GFP-ZO1) during zippering.

How do actin rings form? It has been assumed that these structures are inherited during cell division. By contrast, imaging cell divisions in intact live embryos showed that these structures disassemble prior to division and form de novo after cytokinesis. The mechanism forming the rings after division requires the combined action of cortical flows (from the nascent junction to the cell pole) and a special network of polar microtubules that clears F-actin from the cell pole.

VIDEO.  A network of polar microtubules clears F-actin to form the actin ring de novo after division.

Mechanism for blastocyst formation. Actin rings form de novo after cell division, expand over the entire apical cortex, and zipper along the cell-cell junctions to seal the embryo and enable expansion of the blastocyst cavity (Zenker et al, 2018, Cell).

5. Keratins as a new form of asymmetrically inherited fate determinant

We also recently imaged keratin filaments in the early embryo. These intermediate filaments form a beautiful cytoskeletal network in the mouse and human blastocyst. But unlike actin and microtubules, their function during early development remained elusive (Lim et al, 2020, Nature).

We found that a subset of cells of the mouse and human embryo forms these keratin filaments before the others, even before the establishment of the first differentiated cell lineages (Lim et al, 2020, Nature).

Using high-resolution imaging directly in the living embryo revealed that keratins become asymmetrically inherited by outer cells during cell division. Moreover, the cells inheriting keratins are specified to a trophectoderm lineage fate (ie, future placenta) (Lim et al, 2020, Nature).

The images show fluorescentely-labeled keratins as they become asymmetrically inherited within an intact-developing embryo.