Upload
vantram
View
215
Download
0
Embed Size (px)
Citation preview
AIX-MARSEILLE UNIVERSITE
FACULTE DES SCIENCES DE LUMINY
PH.D THESIS
Presented for defence on 20th January 2015 by
KAPIL BAMBARDEKAR
to obtain the degree of
Doctor of Philosophy (Ph.D.)
of the Aix-Marseille Université Specialization : Biophysics
APPLYING OPTICAL TWEEZERS IN VIVO:
A BIOPHYSICAL STUDY OF MECHANICAL
FORCES IN DROSOPHILA MELANOGASTER
AT THE ONSET OF GASTRULATION
Thesis directed by
PIERRE-FRANCOIS LENNE
in the Insitute of Developmental Biology Marseille Luminy (IBDML)
Referees: Emmanuel Courtade
Sylvie Dufour
Examiners: Hugues Giovanni
Atef Asnacios
Director: PIERRE FRANCOIS LENNE
SUMMARY
The goal of this thesis was to understand the nature and polarization of mechanical forces in
morphogenesis, for which the Drosophila embryonic epithelium was chosen as a model system. The
mechanical properties of cells and tissues have long been thought to play a crucial role in morphogenesis .
However there is a paucity of tools available to quantify and accurately measure mechanical forces in
vivo. Of all the approaches to understand mechanical forces, optical tweezers have shown the most
promise. Here, an optical tweezers setup was developed on a pre-existing single-plane illumination
(SPIM) setup. It was observed for the first time, that the cell-cell interface in embryonic epithelia could be
trapped and manipulated directly with optical tweezers. The interaction of the interface with the trap was
initially characterized at the end of cellularization where the tissue has minimal movements and acto-
myosin turnover. With a sinusoidal trap excursion, the interface amplitude was found to increase linearly
with applied laser power as well as trap amplitude and time period. To correlate the interface movement
to its tension, the trap stiffness was estimated in vivo with injected polystyrene beads. Furthermore, push
and pull experiments on the interface responding to a stationary trap, provided another way to address the
viscoelastic properties of the interface. The interface kinetics in stationary experiments could fit
adequately to a passive viscoelastic model. This model also explained well the linear response to trap
amplitude and time period, and formed the basis of estimating interface tension from its amplitude.
Moreover, the propagation of the sinusoidal movement to neighbouring interfaces decayed rapidly with
minimal phase lag in both experiments and the model. Having established a suitable regime of trapping
conditions, where interface deflection is small and linear, the mechanical anisotropy of the epithelium was
at the onset of gastrulation (early germband elongation). The interface tension increased by 2-3 fold
compared to end of cellularization, exhibiting both apico-basal and dorso-ventral polarization of tension,
concomitant with polarized accumulation of myosin. The role of myosin was established further through
ROCK-inhibition, which restored the tension. Perturbation of actin also decreased the interface tension,
additionally displaying creep-like behaviour. My work provides a crucial insight into the mechanical
behaviour of dynamic epithelia, as well as developing further the field of in vivo optical manipulation.
La mécanique des cellules et des tissus joue un rôle crucial durant la morphogenèse. Cependant, les outils disponibles pour mesurer les forces in vivo sont très rares. Parmi les approches potentielles, nous avons retenu les pinces optiques qui ont été largement utilisées sur des systèmes in vitro, molécules individuelles ou cellules isolées. Nous avons développé un dispositif combinant pinces optiques et imagerie par feuillet de lumière. Nous montrons que les interfaces cellulaires de l’épithélium précoce de l’embryon de Drosophile peuvent être piégées et manipulées directement avec des pinces optiques. Dans un premier temps, la manipulation optique est réalisée à la fin de la cellularisation, processus par lequel des membranes cellulaires séparent les noyaux pour donner naissance à un épithélium ; à ce stade, les mouvements cellulaires sont minimes et les cellules ont des formes hexagonales similaires. En imposant un mouvement sinusoïdal au piège perpendiculairement à une interface, nous étudions la déflection de l’interface en fonction de la puissance laser, de l’amplitude du mouvement du piège et de la fréquence d’oscillation. La réponse est linéaire pour les petites déformations (amplitude de déflection de l’interface inférieure à 500 nm). Afin d’extraire des valeurs de tension, nous estimons la raideur du piège en comparant la déflection imposée directement par le laser à celle produite par des billes individuelles piégées forçant le déplacement de l’interface. Les tensions mesurées sont de l’ordre de quelques dizaines à quelques centaines de pN. En outre, des expériences de déflection–relaxation par déplacement instantané puis arrêt du piégeage, ont été réalisées, fournissant une alternative à l’analyse fréquentielle pour étudier les propriétés viscoélastiques de l’interface. Un modèle de type solide linéaire standard rend compte des observations et permet d’extraire les paramètres viscoélastiques de l’interface. Nous mettons également en évidence que la déflection imposée à une interface se propage aux interfaces voisines en s’affaiblissant exponentiellement sur une distance d’une à deux cellules. Cette technique étant établie, nous l’utilisons pour mesurer les tensions durant l’extension de la bandelette germinale. Les tensions sont anisotropes, les jonctions parallèles à la direction dorsoventrale ayant une tension trois fois plus élevée que celles perpendiculaires. Nous mesurons également des tensions aux interfaces cellulaires plus grandes dans le plan des jonctions adhérentes que dans les plans plus basaux de l’épithélium. Ces tensions sont significativement réduites par inhibition de l’activité du moteur moléculaire Myosine-II. Ce travail fournit pour la première fois des mesures absolues des tensions intercellulaire et un outil pour l’étude quantitative de la mécanique épithéliale in vivo.
CONTENTS
CHAPTER 1 : INTRODUCTION ............................................................. 1
1.1 OVERVIEW.………...………………………………………….......... 3
1.2 ORGANIZATION OF THE TISSUE AND UNDERLYING MECHANICAL ELEMENTS
………………………………………………………….……………….. 5
1.3 CURRENT MECHANICAL UNDERSTANDING OF SINGLE CELLS AND TISSUES IN VITRO ………..……………………………….…………………………..9
1.4 MORPHOGENESIS: TISSUE DYNAMICS DURING EMBRYONIC DEVELOPMENT
…………………………………………..……….................................... 13
1.5 MECHANICAL CHANGES UNDERLYING MORPHOGENETIC MOVEMENTS .... 17
1.6 IN VIVO TOOLS FOR DIRECT UNDERSTANDING OF FORCES ……….......... 21
1.7 GERMBAND ELONGATION IN DROSOPHILA MELANOGASTER AS A MODEL
SYSTEM FOR UNDERSTANDING THE MECHANICAL FORCES IN MORPHOGENESIS
……........................................................................................................... 25
1.8 SUMMARY………………………………………............................. 27
CHAPTER 2 : MATERIALS AND METHODS………........................ 29
2.1 MATERIALS………………………………………………………... 30
2.2 SAMPLE PREPARATION…………………………………………. 30
2.3 EXPERIMENTAL SETUP………………………………………….. 31
2.4 OPTICAL TWEEZERS EXPERIMENTS…………………………... 32
2.5 DATA ANALYSIS………………………………………………….. 34
2.6 EXPERIMENTAL LIMITATIONS…………………………………. 34
2.7 QUANTIFICATION OF E-CADHERIN, MYOSIN II AND LIFEACT.36
CHAPTER 3 : ESTABLISHING OPTICAL TWEEZERS AS AN IN VIVO TOOL TO UNDERSTAND CELL-CELL INTERFACE
MECHANICS ....................................................................................... 37
3.1 MOTIVATION……………………………………………………... 39
3.2 RESULTS………………………………………………………….. 41
3.2.1 Characterizing the deflection of cell-cell interfaces imposed by
optical tweezers………………………………………………….…... 41
3.2.2 Dynamics of cell-cell interface interaction with optical tweezers
as a measure of position offset and phase lag………………………. 45
3.2.3 Relaxation of interface deformation ..…………………….…… 47
3.2.4 Propagation of interface deformation ..…………………….…… 47
3.3 DISCUSSIONS……………………………………………………… 50
3.3.1 Origin and nature of in vivo optical forces…………………… 50
3.3.2 Towards a mechanistic understanding of the experiments….. 53
3.3.3 Developing a mechanical model…………………………….. 55
3.4 SUMMARY….……………………………………………………… 58
CHAPTER 4 : NATURE AND ORIGIN OF MECHANICAL FORCES
IN DROSOPHILA GERMBAND ELONGATION ................................ 59
4.1 MOTIVATION……………………………………………………… 61
4.2 RESULTS…………………………………………………………… 63
4.2.1 Probing the anisotropy of mechanical forces in early gastrulation.. 63
4.2.2 Effect of actin perturbation on interface tension…………..….. 68
4.2.3 Temporal and spatial dynamics: Position offset and phase lag
measurements in different stages and perturbations ………………… 72
4.3 DISCUSSIONS……………………………………………………… 74
4.3.1 Interface tension is regulated by myosin-II activity……….… 74
4.3.2 Actin cortex is responsible for interface tension………….. …. 75
4.3.3 Creep and sinuosoidal loading………………..……………….. 76
3.4 SUMMARY….……………………………………………………… 77
5 CONCLUSIONS AND PERSPECTIVES ............................. ……… 79
ACKNOWLEDGEMENTS .....................................................................................I
REFERENCES .......................................................................................................... II
MANUSCRIPT (IN PRESS PNAS)
LIST OF FIGURES
Figure 1.1 Organization of the tissue…………………………………………… 4
Figure 1.2 Mechanical elements in cell and tissue organization………………. 6
Figure 1.3 Terminology and methods to understand cell and tissue
mechanics………………………………………………………………………... 8
Figure 1.4 Mechanical models for cells…………...……………………………10
Figure 1.5 Cell movements and rearrangements in development…………….. 12
Figure 1.6 Actomyosin dynamics in morphogenesis………………………...….14
Figure 1.7 Measuring mechanical forces in morphogenesis…………….……... 16
Figure 1.8 Mechanical models of morphogenesis…………………………..... 18
Figure 1.9 Methods for in vivo force measurements…….…………………... 20 Figure 1.10 Overview of optical tweezers and particle tracking rheology….. 22
Figure 1.11 Germband elongation: molecular and cellular events…..……... 24
Figure 1.12 Mechanical forces during germband elongation.................……... 26
Figure 2.1 Optical setup combining light sheet microscopy and optical
tweezers…………………………………………………………………………. 31
Figure 2.2 Calibration of trap power and displacement……………………... 32
Figure 2.3 Noise in interface position detection…………………………….. 35
Figure 3.1 Characterizing the deflection of cell-cell interfaces imposed
by optical tweezers……………………………………………………………. 40
Figure 3.2 Interface deflection as a function of trap movement
amplitude and time period of trap movement oscillation…………………... 42
Figure 3.3 Offset between laser and interface positions……………………. 44
Figure 3.4 Phase lag between laser and interface movement…………………46
Figure 3.5 Mechanical model of the interface and tissue response …………... 48
Figure 3.6 Interface deflection induced by cytoplasmic trap………………… 49
Figure 3.7 Quantitative Phase Microscopy Image obtained on early
Drosophila embryo ………………………………………………………….…50
Figure 3.8 Comparison of interface deformation with and without beads… 51
Figure 3.9 Viscosity measurements obtained from bead trajectories………… 52
Figure 3.10 Relaxation of interfaces after trap release. Initial speed of
relaxation is related to the friction coefficient………………………………….54
Figure 3.11 Elongation of interfaces adjacent to the optically deformed
interface…………………………………………………………………………55
Figure 4.1 Tension at cell contacts before and during germband
elongation in normal embryos…………………………………………………. 62
Figure 4.2 Polarity of interface tension at different stages for interface
orientation along the dorso-ventral plane……………………………………… 64
Figure 4.3 Apicobasal polarity of interface tension at different stages……... 65
Figure 4.4 Apicobasal polarity in the lateral plane at the end of
cellularization……………………………………………………………………66
Figure 4.5 Effect of myosin II perturbation on interface tension…………… 67
Figure 4.6 Influence of actin intensity at different stages on interface
tension………………………………………………………………………….. 69
Figure 4.7 Effect of actin perturbation on interface tension………………… 70
Figure 4.8 Phase lag and offset in different embryos ………………………… 71
Figure 4.9 Time dependence and creep after cytochalsin treatment…………. 73
LIST OF ABBREVIATIONS
ECM – extracellular matrix
FA – focal adhesions
AFM – atomic force microscopy
FRET – forster resonance energy transfer
AOTF – acousto-optical tunable filter
NA – numerical aperture
SPIM – selective plane illumination microscopy
GUI – graphical user interface
ROCK – rho-associated protein kinase
PSF – point spread function
– interface sinusoidal amplitude
– trap sinusoidal amplitude - trap stiffness
ini_off - initial offset between trap and interface
fin_off – final offset between trap and interface
int_mov - interface movement when trap is switched on
SD – standard deviation of mean
– trapping force
T – interface tension
l0 – initial junction length
– trap position
x – interface position - interface friction coefficient
SLS – standard linear solid
A/P – anterio-posterior
D/V – dorso-ventral
lat A – latrunculin A
cyto D – cytochalasin D
WT – not injected / control
CHAPTER 1
INTRODUCTION
1
2
1.1 Motivation
Over the last few decades, significant progress has been made in understanding single cell mechanics.
Because cells can be isolated and cultured in vivo, biological mechanisms such as cell migration,
adhesion as well cell-division, tissue growth can be studied from a reductionist perspective. Such an
approach has also facilitated the development of techniques to manipulate and probe mechanical forces in
these events. Particularly, this has led to a detailed understanding in the role of the acto-myosin
cytoskeleton in imparting dynamic mechanical properties to the cell which is crucial in developing and
maintaining an equilibrium of forces with the extracellular environment. The rich diversity of mechanical
models that has developed has also helped in extending this knowledge at the tissue level. In recent years
a lot of focus has shifted in understanding these properties in the natural tissue environment. Here I give
an overview of current understanding in the field and explain the background and context of my work.
3
Figure 1.1 Organization of the tissue
a) A schematic of the typical organization of the cytoskeletal filaments inside an eukaryotic cell.
b) Cell–cell and cell–ECM junctions in epithelia. (a Pullarkat et al. 2007, b Rodriguez et al. 2013)
4
1.2 Organization of the tissue and underlying mechanical elements
The typical organization of the eukaryotic cell is illustrated in Fig.1.1a. Briefly it is enclosed in a bilipid
membrane with various organelles embedded in an viscous cytoplasm. The nucleus and organelles carry
out their functions and maintain their organization by means of a cytoskeleton. Different filaments
constitute to the cytoskeleton, mainly actin filaments, microtubules and intermediate filaments which are
distributed throughout the cell in an organized manner (Fig. 1.2a). The actin filaments form a meshwork
(cortex) below the plasma membrane, and they are cross-linked by a variety of proteins, including motor
proteins, which are capable of generating forces and performing mechanical work. The filaments connect
and interact with the plasma membrane via trans-membrane proteins embedded in the lipid bilayer. The
microtubules originate from the centrosome, close to the nucleus, extending to the periphery, at the actin
cortex. Intermediate filaments are concentrated around the nucleus and extend away from it. The actin
network is the most important cytoskeletal entity for mechanical functions like control of cell shape and
cell locomotion. They are capable of rapid reorganization through de-polymerization and polymerization
cycles and can generate active contractile forces with the help of motor proteins (primarily different types
of myosin). The hydrolysis of Adenosine triphosphate (ATP) to Adenosine diphosphate (ADP) provides
continuous energy for the system. A single reaction liberates about 10 kT of free energy per molecule.
(Alberts et al. 2002, Bray et al. 2000, Pullarkat et al. 2007)
Acto-myosin complexes are formed by the association of myosin motors with actin filaments. Myosin-II
self-assembles into short bipolar chains, which act on neighbouring actin filaments to produce relative
motion and active stresses within the actin network. Acto-myosin complexes can be highly organized, for
example in stress fibers or myofibrils, which consist of contractile bundles of actin filaments and myosin
motors. They also exist as a random network of highly cross-linked “active gel” with mesh sizes of ~100
nm. A good example is the actin cortex, which forms a thin layer (∼ 1 µm) attached to the plasma
membrane in eukaryotic cells. The cortex is the main cytoskeletal component responsible for a range of
functions including control of cell shape, generation of active stresses and cell locomotion. (Alberts et al.
2002, Bray et al. 2000) In all of these cases, contractility results from the activity of myosin II motors
associated with bundles of actin filaments. Myosin motors attach and detach from actin filaments in a
cyclic manner and while attached they undergo a conformational change tightly coupled to ATP hydrolysis
that moves the motor along the filament and generates a displacement of the filament. Because of their
non-processivity and low duty ratio, single Myosin II motors spend an appreciable fraction of their time
detached from the filament. As a consequence, a single motor is not sufficient for motility but their
assembly into mini-filaments converts them into a highly processive motor complex. (Gorfinkiel et al.
2011)5
Figure 1.2 Mechanical elements in cell and tissue organization. a) A schematic showing the coupling
between the cytoskeleton and the extracellular matrix. Cell adhesion is mediated by specialised adhesion
proteins which, typically, form complexes called focal adhesions. The actin cytoskeleton is mechanically
coupled to the substrate at the focal adhesions. (Pullarkat et al. 2007) b) Schematic representation of how
the protein composition of focal adhesions (FAs) is re-organized in response to mechanical force. Within
immature FAs, force-insensitive proteins (grey squares), force-sensitive proteins (blue shapes) and force-
responsive proteins (green shapes) coordinately transmit the specific integrin-mediated signals. In response
to mechanical force, focal adhesion abundance of force sensitive proteins (blue shapes) and force-
responsive proteins (green shapes) are decreased, while the abundance of force-sensitive proteins (orange
shapes) and force-responsive proteins (yellow shapes) are increased. (Kuo 2013). c) Schematic of E-
cadherin adherens junction in epithelial cells. The extracellular domains of E-cadherin homodimers enter
homotypic Ca2+-dependent-binding interactions with those of dimers on adjacent cells. The intracellular
domain of E-cadherin interacts with the actin cytoskeleton via α-catenin and either β- or γ-catenin.
Cadherin–catenin complexes constitute the adherens junction. (Perry et al. 2010) d) Cadherin–actin
interaction through a-catenin .The structure and functional domains of a-catenin in an ‘open’ form,
showing b-catenin, vinculin and F-actin (actin filament) binding regions. The central domain can mask the
vinculin-binding site and can unmask this site when the C-terminus is pulled by acto-myosin forces. Below
- Simplified classical model of F-actin linkage at the AJ. (Yonemura 2011)
6
Cells organize together to form tissues in vivo (Fig. 1.1b, illustrates an example of tissue (endothelium)
with various organizational elements.). Within tissues (as well as in monolayers cultured on artificial
substrates) cells exist in close association with an extracellular matrix (Fig. 1.2a,b). Several cell types
cannot proceed through normal cell division cycle if not attached to this matrix by trans-membrane
proteins belonging to the class of integrins which form mechanical connections with bundles of actin
filaments to form stress fibers. Focal adhesions are sensitive to cell surface chemistry and are promoted
by proteins like fibronectin, collagen etc. Focal adhesions anchor the cell and its actin skeleton to
the substrate and also perform important signaling roles. Interactions between focal adhesion proteins
and extracellular matrix can regulate the cytoskeletal structure as well as active contractile state of the
cell. (Burridge et al. 1996, Zhu et al. 2000)
In addition cells associate with each other through different types of cell- cell junctions. I will focus here
on cell-adherens junctions which regulate adhesion and force equilibrium between cells in a tissue
(Ladoux et al. 2000). Adherens junctions consist of cadherin-class trans-membrane proteins that associate
through homophilic adhesions to form clusters that adhere cells together (Fig. 1.2c, Perry et al. 2010).
Based on ultrastructural observations, the AJs are characterized as a region at the interface of two adjacent
cells with opposing membranes typically ~20 nm apart, with an intercellular space spanned by molecular
strands, and with a dense undercoat associated with actin filaments at the cytoplasmic surface. There are
several types of AJs. Typically, in highly polarized epithelial cells, the AJ encircles the cell completely at
the apical/ basolateral border like a belt and is called the zonula adherens (ZA). Close to the ZA, the tight
junction (TJ) forms apically. Punctate forms of AJs are called punctum adherens (PA). Homophilic
binding results in recruitment of ARP2/3 complex (responsible for branched actin
polymerization) to the adhesive contact (Yonemura 2011). In cadherin junctions (zonula adherens), the
intracellular domain of cadherin associates with the actin cortex through accessory proteins (mainly
catenins). Cadherin junctions are crucial for cells to adhere, migrate, segregate and differentiate in a
selective and coordinated fashion. Moreover they play an important mechanical role in regulating and
responding to cell –cell interface tension (Fig. 1.2d). At the cellular level, the behavior of the cadherin–
catenin complex depends on a-catenin and acto-myosin (Fig. 1.2d). Binding of cadherin associated a-
catenin to actin can occur through interaction with actin-binding proteins, but also directly through its
actin-binding site in a stretch dependent manner. (Yonemura 2011).
The mechanical and active behaviour of cells are surprisingly different when probed in suspension and in
contact with specifically treated substrates which will be reviewed in the following sections.
7
Figure 1.3 Terminology and methods to understand cell and tissue mechanics. a) Basic mechanical
terms (i) Diagrams illustrating definitions of tension, compression and shear forces on a block of
material. Open arrows indicate directions of forces on the block before (left) and after (right) deformation.
(ii-iii) Graphs showing the relationship between stress and strain after application of step changes in stress
(ii) or after rapid changes in strain (iii). The responses of four hypothetical materials are shown: elastic
(an ideal spring), viscous (a fluid), and two examples of viscoelastic materials. Also shown in ii are
simple network models in which the four materials are represented by combinations of springs and
dashpots. (Davidson et al. 2009). b) Viscous phase lag in sinusoidal stress. c) A schematic showing the
main rheometry techniques by which the viscoelastic propertied can be probed at different length and
time scales. The double arrows indicate the sense of applied deformations or force. Laser light is shown in
red. (Pullarkat et al. 2007)
8
1.3 Current mechanical understanding of single cells and tissues in vitro
Typically three schemes are followed to probe the viscoelastic properties of soft materials. These are (i)
creep experiments, where a step force is applied and held constant ,the resulting strain is measured as a
function of time, (ii) relaxation experiments, where step strain is applied and the resulting relaxation of
force is measured, and (iii) oscillatory experiments, where a sinusoidal strain is imposed on the cell and
the phase as well as the amplitude of the force response is measured (Fig. 1.3a and b). In practice, the
coupling between mechanical stress and strain is empirically determined by experiments on how materials
deform in response to applied forces. “Stiffness” in such experiments refers to the elastic response of a
body to an applied force and depends on the its geometry and material properties. (Davidson et al. 2009).
Various techniques have been developed to probe cell mechanics (Fig. 1.3c).Traction force microscopy is
a mostly passive technique aimed at measuring the active force generated by cells in response to different
mechanical or chemical perturbations or in mapping the traction forces generated by locomoting cells. It is
an approach with minimal perturbation to the cell, which is allowed to take any shape it would like.
Initially elastic substrates were used and the stress field was studied by looking at substrate wrinkles
caused by the active contraction of the cells. A major improvement in the methodology are non-wrinkling
silicon substrates where the deformation field can be mapped by measuring the displacement of embedded
beads. A much more sophisticated procedure are arrays of discrete, sub-micrometer-sized elastic pillars,
resembling a fakir bed, on which the cells can adhere and crawl. Micropipette aspiration is a relatively
simple technique where a micropipette with a tip diameter of a few micrometers is used to aspiratea
portion of a cell by applying a known negative pressure. Micropipettes have diameters of about 1–8 µm,
suction pressures are about 0.1–105 Pa. This technique has been successfully applied to probe the
mechanical properties of cells like the red blood cells which have a simple thin spectrin cytoskeleton
tightly adhering to the membrane. Atomic force microscopy (AFM, Fig. 1.3c ii) can probe cell mechanical
properties at a submicrometer scale. Typically, forces in the range of 0.1–1 nN and indentations of about
50 nm or less are applied to the cell to probe its viscoelastic response. Frequency responses of both storage
and loss moduli in the range of 0.1–300 Hz can be measured if corrections to hydrodynamic artifacts are
applied. Modified AFM tips, where a bead with a diameter of a few microns is glued to the normally sharp
tip, allows the modification of the stress range by changing the bead radius and diminishes the
indentations. Magnetic bead microrheology (Fig 1.3c iii) generates well defined forces on tiny
paramagnetic beads with an electromagnet to probe the mechanical response of cells. The beads can be
attached to the outside of the cell using specific adhesion promoting proteins to form transmembrane
9
Figure 1.4 Mechanical models for cells. a) Overview of different mechanical models (Lim et al. 2006).
b) Steady-state contractile force as a function of support stiffness predicted with the constrained mixture
model. The steady-state morphology and corresponding stress fiber distributions are shown for select
values of substrate stiffness. For comparison, experimental results from Ghibaudo et al. 2009 are
also reported. (Rodriguez et al. 2013). c) Model of cadherin contact formation and strengthening in
response to mechanical changes in the cell-cell contacts. The substrate is coated with N-cadherin and
represents a neighbouring cell. The close-ups of cadherin contacts show the balance of external and
internal forces (Fext and Fcell, respectively). (Ladoux et al. 2010)
10
By combining experiments and modeling detailed understanding of cell mechanics in various in
vitro systems has been deduced. Fig. 1.4b shows comparison of results obtained by traction force
microscopy and then constrained mixture model. This model is able to accurately capture the relationship
between cell contraction and substrate stiffness and predicts correctly, the formation and orientation of
stress fibers in cells stretched on substrates with different stiffness. (Rodriguez et al. 2013, Fig. 1.4b).
Fig. 1.4c illustrates an experimental model of cadherin contact formation and force response, deduced
from traction force microscopy experiments. As the cell pulls on the substrate via cadherin adhesions, it
induces an increase of its internal tension by recruiting adhesion proteins and upregulation of acto-
myosin contractility. On a stiff substrate (large K), the internal tension (Kint) is supported by the
formation of large clusters of cadherin complexes. When the cellular environment provides less
resistance to deformation (small k), small forces are observed with limited cadherin links and
therefore, smaller internal rigidity (Kint). (Ladoux et al. 2010 , Fig. 1.4c).
11
linkages with the cytoskeleton (Pullarkat et al. 2007). Techniques like micropipette aspiration can effectively probe the creep response of cells, while AFM, magnetic twisting cytometry, microplates, etc, can perform oscillatory probing at different frequencies. AFM and magnetic twisting cytometry offers the widest range of frequencies, from 0.01 Hz up to 1 kHz. The microplate technique (iv), although limited in frequency, can perform variety of measurements not easily achievable in other methods. On the other hand, optical techniques like laser tweezers (i) and optical stretcher (v) can make rheological measurements on non-adhering cells. (Pullarkat et al. 2007).
Mechanical models for biological materials are derived using either the micro/nanostructural approach or the continuum approach. The micro-structural approach focuses on the cytoskeleton as the major structural component and is applied especially for understanding cytoskeletal mechanics in adherent as well as floating cells.. The continuum approach attributes continuum material properties to cell components which is more straightforward for understanding the biomechanical response at the cell level. It also provides a distribution of stress and strain on the cell which can be used to calculate transmission and distribution of forces/tension. This can furthermore assist in the development of more accurate micro and nanostructural models. (Lim et al. 2006, Fig. 1.4a).
Figure 1.5 a) Tissue movements and rearrangements during development. (i) Mesoderm invagination
on the ventral side of the Drosophila embryo starts 3 h after fertilization and is completed within 5 min.
Presumptive mesoderm cells undergo pulsatile apical contraction as they start the process of gastrulation.
Right; graph showing the stepwise cell area reduction of these cells and the fluorescence intensity levels of
a MyosinGFP reporter. (ii) Dorsal closure in Drosophila embryo. This process takes approximately
three hours from the end of germ-band retraction with slow AS contraction starting after around 45 min.
Right; anticorrelation between the cell area and MyosinGFP intensity levels in amnioserosa cells.
(Gorfinkiel et al. 2014). b) Cell reorganization and shape changes underlying tissue movements
i) convergence of single cell boundaries ii) Contraction of multicellular boundaries iii) Apical constriction
and iv) directional basal constriction (Baum et al. 2011).
12
1.4 Morphogenesis: Molecular events and dynamics during embryonic
development
Development of a single-celled embryo into a complete organism is a complex process involved many
changes and dynamics at the molecular, cellular and tissue level. Cells form tissues (epithelia) through
cell division. With the onset of gastrulation cells undergo variety of changes in their morphology/shape
and also remodel their contacts. This process of tissue morphogenesis has been studied for many years in
the context of biochemistry and molecular players involved. But with the advent of time lapse imaging,
lot of focus has been shed recently on the temporal and spatial dynamics of this process using
various model systems (Gorfinkiel et al. 2014). Initially reported in the C. elegans oocyte, thus far,
epithelial cells in at least five tissues undergoing morphogenesis have been shown to exhibit episodic
actomyosin behaviour, four in Drosophila and one in the Xenopus embryo. Mesoderm invagination and
dorsal closure during Drosophila gastrulation are 2 such examples. Mesoderm invagination is the first
step in Drosophila gastrulation and leads to the formation of the mesoderm. Dorsal closure occurs much
later wherein the dorsal amnioserosa undergoes contraction and disappears while the surrounding
tissues closes the gap (which has also been used as a model for wound healing). Both these
processes exhibit pulsatile behavior in cell changes (apical constriction for cells in mesoderm
invagination, contraction of cell area for amnioserosa) (Fig. 1.5a). Such pulsatile behavior for cell
reorganization and shape changes has been observed in other model systems as well. Fig. 1.5b shows a
schematic of the various shape changes that can be observed in tissue morphogenesis (Baum et al. 2014).
Cells in an epithelium can shrink or contract their boundaries leading to loss of junctions (cell
intercalation, results in formation of now boundaries/junction in a perpendicular direction, eg.
germband elongation) (i). Cells can also coordinate together to contract multicellular boundaries,
forming rosette-like structures (ii) as well as constrict their apical and basal ends (Fig. 1.5b iii and iv).
Cells actively produce these shape changes by generating forces at their surfaces, which are transmitted
through cell interfaces. These ‘cortical’ forces, are generated in the cell cortex. Cortical forces build up
from a variety of mechanisms, importantly from Myosin II and actin filaments assembly which can be
spatially and temporally controlled in the cell. Understanding how these forces emerge from the assembly
and contraction of acto-myosin networks coupled to adhesion structures is a central issue in cell and
developmental biology. Therefore the spatiotemporal behavior of acto-myosin has been studied in
many systems. Indeed in some of these systems pulsatile changes in actin and myosin levels are
observed with the onset of cell contractions (Fig. 1.6b). The frequency of actomyosin foci and of the
13
Figure 1.6 Acto-myosin dynamics in morphogenesis. (a) Schematic representations of the acto-
myosin organization and ratchet mechanisms within cells of the model tissues (red; actin, green, myosin).
Cell views are apical. In the Drosophila mesoderm cells, high frequency myosin foci reinforce themselves
and contribute to the progressive development of an apical ratchet as the acto-myosin network becomes
connected across cells. Both junctional recruitment and the development of a multi-cellular network of
acto-myosin as the frequency of foci increases contribute to increasing stiffness of the amnioserosa. (b)
Time offsets of the maximal rate of change in cell area (or junction length for the germ-band cells)
relative to the maximal rate of increase in fluorescence intensity of myosin and/or actin. (c) Range of
fluctuation period lengths in all model tissues. (Ect, ectoderm; EM, early mesoderm; LM, late mesoderm).
(d) Tissue contraction rate versus fluctuation period length (DMI, Drosophila mesoderm
invagination; GB, germ-band; AS, amnioserosa; ND, not detectable). (Gorfinkiel et al. 2014)
14
resulting fluctuating contractility of cells in the model tissues is similar The pulses of acto-myosin
changes as well as well cell contraction are coordinated, and seem ot occur between intervals of a few
minutes (1-10 minutes, Fig. 1.6c). The amount of tissue contraction resulting from coordinated cell
contractions shows a correlation with the period of fluctuation/pulses (Fig. 1.6d). In the absence of
intrinsic cellular mechanisms that generate polarity in the plane of the epithelium, anisotropic tissue
tension could organise the spatial architecture of the actomyosin cytoskeleton and orient the flow of
foci. Both mesoderm and amnioserosa cells in Drosophila do not exhibit planar localization of junctional
proteins but experience anisotropic tension that is greatest along the A/P axis. Pulsatile actomyosin
contractility within cells raises the question of how tissues effectively contract. A ratchet
mechanism has been postulated for the stabilization of fluctuations and the generation of net contraction.
Both medial and junctional actomyosin populations contribute to tissue remodelling in various
tissues. In the amnioserosa, an increase in junctional actomyosin fluorescence is observed during dorsal
closure and is correlated with a straightening of initially wiggly membranes and productive tissue
contraction. In the germ-band the dynamic apical foci are thought to feed and reinforce the D/V oriented
junctions. In both these tissues it is possible that the junctional population acts as a ratchet that maintains
the contraction generated by the medial actomyosin network. Overall this suggests that cell contractions
are regulated by acto-myosin ratchets (Fig. 1.6a illustrates that the organization of acto-myosin and
the ratchets involved in mesoderm invagination and dorsal closure). (Gorfinkiel et al. 2014, Lecuit et al.
2011, Rauzi et al. 2011).
Elucidating the forces that form and reshape multicellular structures is thus integral to the understanding
of development. It is therefore important to understand the material properties of in vivo epithelia
to understand how dynamic acto-myosin behavior evolves the forces within the tissue.
15
Figure 1.7 Measuring mechanical forces in morphogenesis. a) An example of laser microsurgery that cuts
a single cell-cell boundary . The cut (blue arrow in 0sec) locally releases tension, which creates an opening
in the tissue. The cells surrounding the opening (in red) and cells further away (in green) are analyzed in
80sec and 180sec. (Solon et al. 2009). b) Annular ablation experiment in the scutellum. The epithelial cell
apical junctions are marked by E-cadherin:GFP. The region between two concentric circles indicated by blue
lines defines the annular severed region (left). The circular domain retracts after the cutting as shown in the
images 1 s after (middle) and 30 s after (right) cutting. Yellow: fitted ellipse. The larger opening along y-
than x-axes indicates anisotropic stress in the tissue. (Ishihara et al. 2013). c) Estimated tensions and
pressures for a Drosophila pupal wing. (i) An image of a Drosophila wing at 23 h APF. D_catenin-TagRFP is used to highlight cell shape. Scale bar: 20 μm. (ii) Color maps of estimated cell-junction tension using ST.
A color scale is shown. (iii) Comparison of estimated tensions obtained using STP and ST. (Ishihara et al. 2013).
16
1.5 Mechanical changes underlying morphogenetic movements: Current
methods and theoretical understanding
Understanding the mechanical nature of morphogenetic movements requires force measurements in vivo.
However, while a variety of techniques have been developed for in vitro measurements of cells and
tissues, applying these techniques in vivo is challenging due to the direct mechanical contact required
between the probes and the cell/tissue required in most of these techniques. As cells and tissues within
embryos respond differently to external stimuli compared to isolated cells on artificial substrates, it
requires a unique approach to make these measurements.
Most information on mechanics in embryos comes from indirect methods. A well-studied example is laser
ablation. In this technique a highly focused and high power femtosecond laser beam is pulsed within the
cell or tissue to induce perturbation/breakage. In the context of mechanics, at the single cell level this
typically involves making a ‘cut’ within the acto-myosin cortex by focusing the laser on a spot of
few 100nms (Rauzi et al. 2011). As the cortex is under dynamic tension which is suddenly ‘released’
by the ablation, it tries to relax in response (Fig. 1.7a amnioserosa during dorsal closure, Solon et al.
2009). This can be measured with time-lapse imaging as a function of cortex movement away from
the point of ablation. A variant of this approach is to do a tissue level ablation (for eg. A circular/
annular cut) and measure the relaxation at the tissue level. An example in the Drosophila scutellum is
shown in Fig. 1.7b (note the anisotropic relaxation in the anterior-posterior and dorso-ventral
axes). The nature and anisotropy of tension in epithelia can be estimated. However since the
relaxation depends on the cytoplasmic friction coefficient (for which there are no direct measurement
in most systems), in addition to the tension; it only provides an indirect estimate of the forces involved.
Nevertheless due to its non-invasive nature, laser ablation has been used to understand epithelial
mechanics in various experimental models of morphogenesis including germband elongation, dorsal
closure and mesoderm invagination (Rauzi et al. 2008, Solon et al. 2009).
An alternative to perturbation is force inference from image analysis. The force inference method is non-
invasive and provides space-time maps of stress in a whole tissue, unlike existing methods (Fig. 1.8iii,
example in wing disc). Different force-inference methods differ in their approach of treating
indefiniteness between cell shapes and forces. Tests using artificial and experimental data sets
consistently indicate that the Bayesian force inference, by which cell-junction tensions and cell pressures
are simultaneously estimated, performs best in terms of accuracy and robustness. Moreover, by measuring
17
Figure 1.8 Mechanical models of morphogenesis. a) Experiments and 3D model for Drosophila gastrulation. (i) Ventral (from Grumbling and Strelets, 2006) and cross-sectional (from Muñoz et al., 2007) views of ventral furrow formation in experiments. (ii) Same views from finite element model. b)
Two-step process for invagination in ascidian gastrulation. Step 1: Apical constriction results in wedge-
shaped cells in endoderm surrounded by ectodermal cells. Step 2: Subsequent apico-basal contraction of
endodermal cells results in invagination. c) Cell shape organization in the ommatidium of the Drosophila
retina. The top left panel shows the cell organization in a normal fly, and the bottom left panel is the cell
organization predicted by a simple mechanical model that considered only adhesion energies and
membrane elasticity. The right panels show the cell organization in a mutant fly in which the left cone
cell (black) indicated by the red lines (bottom panel) lacks N-cadherin. The effect of this deletion is
accurately predicted by the mechanical model. (a,b from Wyczalkowski et al. 2012 and c from Niessen
et al. 2011)
18
the stress anisotropy and relaxation, the force inference and the global annular ablation (Fig. 1.8b) of the
tissue can be cross-validated, as each of them relies on different prefactors. (Ishihara et al. 2013)
While experimental probing of in vivo mechanics has only developed recently, mathematical models
have been applied to morphogenesis for a few decades. In one of the earliest theoretical
investigations of morphogenesis, Odell (Odell et al. 1981) presented a 2-dimensional (2D) model for
an epithelium that treats each cell as a viscoelastic element with a contractile apex. In a ring of cells,
contraction in one cell apex (which simulated by shortening the stress-free length) stretches
neighbouring cells, which can also contract themselves if stretched beyond a limit. On selecting the right
parameters this response produces a wave of contraction that generates a local invagination. (similar to
observed in ventral furrow formation and neurulation). Since this significant study, lot of models
have been developed, extending into 3D. Particularly, continuum growth theory based models of
ventral furrow formation can simulate active changes in cell shapes (Fig. 1.8 a and b). In these
models, cell dimensions change by specifying positive or negative growth along particular directions,
while cell wedging occurs via an apico-basal growth gradient. The models were used to study the
effects of various combinations of these cell shape changes, as well as 3-dimensional (3D) ellipsoidal
geometry and constraints imposed by the surrounding vitelline membrane and internal fluid (Fig. 1.8b).
The 3D ellipsoidal model yielded global shape changes similar to those observed in experiments, and
confirmed the important role of the vitelline membrane and yolk in gastrulation. (Bonnet et al. 2012,
Fletcher et al. 2014, Hutson et al. 2003, Munoz et al. 2007, Wyczalkowski et al. 2012,)
Quantitative vertex models can also predict effect of biochemical perturbation on cell-shapes in vivo Fig.
1.8c shows such an example in the Drosophila ommatidium. The cell shapes in the ommatidia are
reminiscent of soap bubbles, whose geometry is determined entirely by surface tension. A simple
mechanical model was sufficient to predict cell geometries in vivo, for different cluster sizes and for
different mutants. But this correlation is only indicated for small cell clusters. In more complex tissues,
other forces might play a significantly greater role to determine cell morphologies than adhesion energies.
(Niessen et al. 2011)
19
Figure 1.9 Methods for in vivo force measurements. a) Schematic of magnetic tweezer perturbation in Drosophila embryos: Tissue level laser ablation and application of external force (1.5 amp for 1.7 min)
using a magnetic tweezers on Drosophila embryo injected with 100 nm paramagnetic beads). (Kumar et al. 2012) b) AFM. (i,ii) Confocal zx profiles of a cell monolayer (green) grown on a soft collagen gel (black),
before (i) and during (ii) indentation with an AFM cantilever (dotted line). White arrowhead, an individual
cell; grey arrowhead, the tip of the cantilever. A fluorescent dye was added to the extracellular medium (red).
Scale bar: 20 mm. (Harris et al. 2014) c) Schematic drawing of the tension-sensing (TS) FRET sensor
module. Teal fluorescent protein (mTFP) is separated from Venus, a yellow fluorescent protein, by a
nanospring protein domain from spider silk. In the relaxed state, the two fluorophores are close enough to
allow FRET. The spider silk domain stretches in response to pico Newton forces, reducing FRET (Grashoff
et al., 2010). d) Force measurments with FRET sensor in Drosophila embryos. (i-iv) Rescue of Armadillo
expression (Arm, which is Drosophila b-catenin) in border cells after EcadRNAi (ii) by CadTS (iii) and
control (iv). Scale bar shows 10 μm. v) Histogram showing CadTS and control rescuing border cell
migration after border cell-specific (ii) EcadRNAi. (Cai et al. 2014).
D
20
1.6 In vivo tools for direct understanding of forces
Most of the information on mechanical forces in vivo comes from indirect approaches. While such
measurements from direct methods are lacking, approaches are nevertheless being developed. In
particular magnetic tweezers have been applied in Drosophila embryos in conjunction with laser ablation,
to probe the effect of mechanical perturbation on morphogenetic movements and the corresponding
gene expression patterns (Fig. 1.9a). Mechanical forces induced by magnetic tweezers in paramagnetic
beads injected in the embryo, altered nuclear morphology as well a induced movement (Kumar et al.
2012). Atomic force microscopy represents another approach to probe-less manipulation. Although it is
limited to probing forces at the surfaces of embryo, it provides the advantage of high-resolution
imaging combined with measurement and application of pN-nN forces. While imaging of live embryos
has been done with AFM, mechanical probing has been limited to tissue monolayers (Fig. 1.9b).
Nevertheless with the advent in technology, it will prove a useful tool for understanding in vivo
mechanics (Harris et al. 2014). FRET-based force sensing is another important approach developed in
recent years (Grashoff et al.). A force sensitive peptide is placed between a FRET donor-acceptor pair.
Depending on the force applied, the distance between the donor and acceptor changes, thus altering
the FRET efficiency (Fig. 1.9c schematic). By calibrating the force versus FRET efficiency in vitro,
the approach can then be applied in vivo. In fact such a measurement has already been done in
Drosophila embryos during border cell migration (Cai et al. 2014, Fig. 1.9d).
Despite these developments, optical tweezers and particle tracking rheology remain the most viable and
least-explored tools to understand morphogenesis. Optical forces can manipulate and apply forces to few
100nm-micron sized probes. The most classical application of this is the single beam gradient optical trap,
also known as optical tweezers (Ashkin 1987). In a tightly focused laser beam, the scattering force that
pushes a probe in the direction of laser propagation is balanced by the gradient force which
is proportional to the intensity gradient of the beam (Fig. 1.10a). When the probed object has a
positive refractive index mismatch with the surrounding medium, it gets ‘trapped’ in the optical
tweezers. Since refractive index variation can arise naturally in biological tissues, native objects such as
lipid vesicles and or organelles can be probed and manipulated. For small displacements up to 1 µm and
for sufficient index mismatch, the trap can be applied with minimal optical damage with a linear force-
dependent spring-like behavior. The stiffness of this spring like system needs to be calibrated for precise
21
Figure 1.10 Overview of optical tweezers and particle tracking rheology a) Origin of Fscat and Fgrad in single beam gradient trap (optical tweezer) for high index sphere displaced from TEM00 beam axis
(Ashkin 1987). b) Manipulation of DNA attached to a glass surface with optical tweezer. (Hormeno et al. 2006). c) Artificial cytoplasmic filaments in a scallion cell. The laser trap is moved from A to B, pulling out the viscoelastic filament AB into the central vacuole (Ashkin et al. 1989). d) An example of in vivo
stall force measurement for motor proteins (kinesin I and cytoplasmic dynein). (left) Snapshots showing a
lipid droplet in an embryo trapped at multiple positions along its trajectory shown in right. (Leidel et al. 2012). e) Schematic for particle-tracking microrheology. Left:The random spontaneous movements
of injected beads within the cytoplasm are monitored with high spatial and temporal resolution.
Right: Typical trajectory of the beads in the cytoplasm. (Wu et al. 2012). f) Left: Inert nanoparticles
injected in C. elegans embryo uniformly dispersed after first cell divison. Right: The bulk diffusion
coefficient of the anterior and posterior cytoplasm. (Daniels et al. 2006)
22
force measurements. In fact, the ability of optical tweezers to achieve probe-less manipulation was
demonstrated long ago (Ashkin 1989, Fig. 1.10c). But classical probes such as silica or polystyrene
microspheres are easier to calibrate for force measurements (Fig.1.10b, Hormeno et al. 2006).
Hence such measurements are traditionally limited to in vitro biochemical systems. However
measurements have been done on motor transport in vivo by trapping endogenous lipid droplets
(Fig. 1.10d, Leidel et al. 2012). Such measurements rely on advances in trap calibration as well
as particle tracking rheology. In particle tracking rheology beads are injected in cells or embryos.
The time-dependent (x, y) coordinates of the beads are mathematically transformed into mean squared
displacements (MSDs). The time lag-dependent MSDs of the beads are subsequently transformed into
local values of either the frequency-dependent viscoelastic moduli or the creep compliance of the
cytoplasm. (Fig. 1.10e, Wu et al. 2012). The measured viscosity depends on the size of the probe, hence
the use of endogenous particles which can vary in size, has not been verified. Nevertheless this
approach has already been used in vivo using injected beads, for example to passively probe the
viscoelastic properties in the C. elegans embryo (Fig. 1.10f, Daniels et al. 2006). Combining passive
particle tracking with active force perturbation with optical tweezers could provide the missing gap
in extending mechanical understanding from in vitro to in vivo, since such active and passive
microrhelogy is well characterized in vitro.
23
Figure 1.11 Germband elongation: Molecular and cellular events a) Cartoon depicting a Drosophilaembryo during gastrulation. The germband (GB) converges in one direction extending in the
perpendicular direction. GB convergence-extension is driven by a cell cell intercalation. Cell intercalation
is polarized along the anterior/posterior axis. (Rauzi et al. 2011) b) Enrichment of Sqh–GFP at type 1
junctions (arrowheads). On left, time-lapse sequence of Sqh–GFP (times in minutes), showing the
enrichment at type 1 junctions (0, pink and orange arrowheads), in type 2 junctions (20 min, pink
arrowheads) and the reduced localization at type 3 junctions (33 min, pink and orange arrowheads).
(Bertet et al. 2004). c) Two-tiered actin model to explain regulation of the stability and mobility of homo-
E-cad clusters by actin, in SAJs. (Cavey et al. 2010). d) Left; localization of Myo-II and E-cad before
and during intercalation. On right; respective distribution of medial (red) and junctional (green) Myo-II
along the apico-basal (z) axis (Rauzi et al. 2010).
24
1.7 Germband elongation in Drosophila as a tool for understanding the
mechanical forces in morphogenesis
The first 13 nuclear divisions of the Drosophila embryo occur in a syncytium, resulting in 6,000
peripheral nuclei located beneath the plasma membrane. During cellularization, the membrane surface
increases ~25-fold, invaginates between the nuclei, and ultimately yields 6,000 epithelial cells 30 μm tall
(Lecuit et al. 2000). The global contraction of the actomyosin network at the leading edge of the furrow
provides the force necessary to pull down the membrane. Myosin enrichment is mostly at the
cellularization front which proceeds basally to complete cellularization. At the onset of gastrulation the
membrane folds disappear and myosin enriches apically. At the onset of gastrulation, the resultant
epidermal cells have the hallmarks of polarized cells with adherens junctions separating the apical and
basal–lateral domains. It is still unknown how this polarity is established. However it has been observed
that cadherin clusters form at the end of cellularization and migrate apically by the time germband
extension occurs (TroungQuang et al. 2013). The Drosophila germband at the onset of gastrulation is an
excellent model system to study cell intercalation. (Fig. 1.11a). It has been suggested that
differential adhesion between groups of cells could drive cell rearrangement (Doubrovinski
et al. 2014), Cells in the germband rearrange their neighbours by remodeling their junction
in a polarized fashion so that junctions parallel to the dorsal/ventral axis (vertical
junctions) shrink bringing four cells in contact and then expand in a direction parallel to the
anterior/posterior axis so that more dorsal and ventral cells form new contacts. (Fig. 1.11b Bertet et
al. 2004) During this process, Myosin II enriches along vertical junctions. Myosin II is
necessary for junction remodeling as the contractile activity of Myosin II might create a local
tension that orients the disassembly of junctions. This hypothesis was tested by a quantitative
comparison between in vivo data and in silico predictions and laser subcellular dissection (Rauzi et
al. 2008). A two-tiered meshanism for regulation of cadherin cluster dynamics has been proposed. In
this model, Stable, small actin patches concentrate and stabilize homo-E-cad in SAJs.
a-Cat is not necessary for stability, and unknown linkers between actin and E-cad are
involved. A dynamic, contractile network regulated by Bitesize and Moesin tethers all SAJs and
limits their lateral mobility through a-Cat (right), thereby maintaining adhesion in a defined domain.
(Cavey et al. 2010).
25
Figure 1.12 Mechanical forces during germband elongation. a) Forces during cell intercalation revealed
by laser nano-dissection. (i) Local ablation of the subcortical acto-myosin network (left) causes
redistribution of F-actin and E-cadherin and changes the force balance at cell junctions (middle). (ii, iii) Time-lapse sequence following nano-dissection, which was performed at t = 0 in a subwavelength
volume (red arrowhead). MoeABD::GFP to mark F-actin (ii), E-cadherin::GFP (iii). Three-photon
uncaging of fluorescein in a single cell ensures that the plasma membrane is not permeabilized.
Kymographs show the temporal evolution of the fluorescence intensity along the targeted junctions
where the nano-dissection was performed. The two vertices of targeted junctions move apart after
nanodissection. Scale bars, 5 μm. (white), 50 s (blue). (iv) νmax as a function of ϕ for v-junctions (vertical) and t-junctions (transverse). Black and red solid curves are fits of νmax for the v- and t-
junctions, respectively. The orange solid curve represents the expected speeds for t-junctions assuming
mechanical equilibrium between t- and v-junctions. (v) Expected tension in a line tension or a cortical
elasticity model for v- and t-junctions in mechanical equilibriumand for a simple geometric
transformation at constant area (inset). (Rauzi et al. 2008). b) Tension anisotropy is sufficient to drive
tissue elongation. (i) A group of cells observed in vivo at the onset and at 40 min of elongation (top).
Comparison with simulations, started from the same cell network, imposing a cortical elasticity anisotropy β = 1.8. Bottom right: in silico states at the same relative elongation (rel. elong.) as that observed in vivo.
Bottom left: in silico final state. (ii) The final relative elongation of the cell network is a function of the
tension anisotropy imposed at the onset of simulations (squares, cortical elasticity anisotropy β). (iii) Tissue elongation as a function of T1 transitions in vivo (red squares) and in silico for different values of
tension anisotropy. (Rauzi et al. 2008).
26
Rauzi et al. (2008) revealed an anisotropy of cortical forces along cell junctions controlled by Myosin II:
this was measured to be a factor 2 along vertical junctions (junctions with greater density of Myosin II)
compared to other junctions. Cortical forces were inferred by laser dissection experiments: disruption of
the acto-myosin network underlying a given junction modified the balance of forces and produced
junction relaxation, whose speed is indicative of cortical tension . The authors designed a model based on
the local (junctional) and global (cellular) natures of cortical forces. The cellular network configurations
during tissue shape changes were described as the succession of local minima of an energy Starting the
simulations with cell patterns observed at the onset of intercalation, we monitored the same cells in silico
and in vivo during elongation (Fig. 1.12a). We studied three embryos for 40 min. Tissue elongation after
complete energy minimization depends on cortical elasticity anisotropy (Fig. 1.12b). For tension
anisotropies below 1.4, the tissue failed to elongate significantly, but above 1.6, final elongation was
maximal, indicating that moderate tension anisotropy is sufficient to drive maximal tissue elongation.
(Rauzi et al. 2008)
27
1.8 Summary
We are at the stage where we continue to develop a better understanding of mechanical equilibrium in
tissues, particularly in the context of cell-cell and cell-ECM adhesion. However recent studies have made
it clear that tissue monolayers in vitro have different properties than in vivo epithelia. Thus it becomes
ever more important to probe tissue dynamics within embryos. Understanding the role of mechanical
forces in tissue morphogenesis is a longstanding problem in biology. Due to advances in
fluorescent labeling and real-time in vivo imaging in recent years, a lot of light has been shed on the
nature of forces in development. Particularly, the dynamic nature of acto-myosin organization in the
form of pulses over a seconds plays a crucial role in morphogenetic events in different model systems.
One such system is the germband elongation in Drosophila gastrulation. Mechanical properties
underlying embryo morphogenesis have been mainly done with through laser ablation and force
inference methods. However methods for direct measurements of forces in vivo are only being
developed now, particularly through advances in particle tracking rheology and optical tweezers.
28
CHAPTER 2
MATERIALS AND METHODS
29
2.1 Materials
Drosophila melanogaster strains with either Gap43::mcherry fluorescent marker ( [w ; UASt-
sqhE20E21/Cyo ; UAS-diaCA/TM6tb II,III], [w, upd/Fm7 ; endocad-GFP, sqh-cherry /Cyo X,II] or [w ;
sqh-GFP::utABD/Cyo ; UASp-Gap43::mCherry/TM6tb II,III]) were used for optical tweezer
experiments. Cadherin imaging was done with endocad::GFP (,[tl1-endoCad-TCSi::GFP-12 II], [w,
upd/Fm7 ; endocad-GFP, sqh-cherry /Cyo X,II]). Myosin imaging was done with squash::GFP (w ; sqh-
GFP::utABD/Cyo ; UASp-Gap43::mCherry/TM6tb II,III) and actin imaging was done with lifeact::GFP
(sqhp-lifeact::eGFP/MKRS II ). Flies were maintained at 22ºC in tubes with fly medium and in a cage at
25ºC for experiments. Fresh medium plates (with yeast) were left for 2:30 hours in the cage before
collecting embryos, to provide a variation of stages and thus enough embryos for experiment (to account
slow experiment with tweezers and rapid development of stages of interest). Alternatively for more
precision, plates were left for 1 hour in cage and then an additional hour at 25C after removing from the
cage. Embryos of interest were typically at the end of cellularization and could be probed till the early
germband elongation. For myosin perturbation 10mM concentration of Y27632 ROCK Inhibitor
(Invitrogen) was used. Actin perturbation was done either with 100 µM cytochalasin D or 1mM
latrunculin A.
2.2 Sample Preparation
To prepare samples for imaging/optical tweezer experiments, plates with embryos in suitable stage were
collected. After washing off yeast from the plate with deionized water, the embryos were washed with
100% bleach for 50 seconds to remove the vitelline membrane, then washed further with water to remove
the bleach and fragments of the vitelline. Embryos in the end of cellularization (stage 5 end) were then
selected under a Zeiss Stereo dissection scope. For deflection experiments, the embryos were glued on the
edge of a custom made SPIM coverslip which was then attached to a custom made chamber and
immersed in deionized water.. For spinning disk and phase measurements experiments, embryos were
aligned in the center of a rectangular cover-glass with a drop of halocarbon oil on top. In every case,
alignment was done with the germband visible in the imaging plane typically with anterior-posterior axis
vertically in reference to the camera image and the dorso-ventral axis horizontal. For perturbation,
embryos were placed in halocarbon oil and injected in a microinjection setup (Eppendorf Femtojet , 250-
400 psi,Ti -0.3, Tc -0) with water, ROCK inhibitor, latrunculin A or cytochalasin D at stage 5 end and
analysed during germband elongation (or the corresponding time point). Injection micro-needles were
custom made from glass capillaries with a glass pulling apparatus. Spilling embryo contents was carefully
30
avoided by adjusting the needle tip to less than cell size (~6 and injecting with a droplet few cell-sizes in
diameter. For perturbation followed by deflection experiments, the halocarbon oil was removed carefully
with a scalpel and rigorous washing under water before imaging.
2.3 Experimental setup
The light sheet setup is similar to the one in the seminal paper of Huisken (Huisken et al. 2004), and is
already well-described in a previous publication (Chardes et al. 2014). But briefly, 3 imaging lasers
(408, 488 and 561nm) are aligned along the same optical path using dichroic mirrors. An AOTF
allows switching the wavelength and power of the outgoing beam. This beam is passed through a beam
expander and cylindrical lens to create a light sheet, which is then focused on the sample using a Nikon
Plan Flour objective (0.3NA, 10x) with a resulting thickness of ~3 µm. Imaging is done with a
Zeiss Axiovert upright microscope. The light from the sample is collected using a 100x water immersion
Figure 2.1 Optical setup combining light sheet microscopy and optical tweezers.
lens (Nikon Plan 1.1NA), passed through dichroic lens and imaged onto a Andor IXON 3EMCCD
camera. Using this setup 3 channel acquisition is possible by alternating the AOTF output
wavelength and also simultaneously by an dual-view system before the camera path that can split the
emission from the GFP and cherry channels on 2 halves of the camera CCD (however only single
channel measurements were used in all the experiments). To enable optical tweezers on this setup, a
31
1071nm infra-red beam was introduced on the optical table using a continuous wave (CW),
Ytterbium Laser Module (YLM) (IPG photonics, maximum output 5W, 3A, 20% efficiency). All
the components for the trap setup were purchased from Thor labs. The beam was first passed
through a polarizer and lambda/4 plate. A system of 2 galvometric mirrors (galvo) was then
introduced in the path. A conjugating telescope (2 50mm lenses) was aligned between the 1st galvo
and 2nd galvo. While keeping the conjugation, the beam was then expanded with a long distance beam
expander to ensure that the objective back aperture was filled and to provide sufficient working
distance. The beam was then lifted to the objective plane using a periscope . Completing the
conjugation and alignment ensured that the laser trap was in focusing plane and the field of view of
the camera. The galvos could be triggered at the same time as the camera acquisition using a
National Instruments data acquisition card and custom software made in Qt creator. After building the
setup, the transmission of the entire setup was calculated with a reflecting Power meter and found to be
20% (fig 2.2 a). The galvo conjugation was confirmed for the experimental conditions by a linear
Figure 2.2 Calibration of trap power and displacement. a) Measurement of trap laser power at
different locations in the optical setup. b) Displacement of trap (as measured using trapped beads in
water) as a function of galvo voltage. c) Measured laser amplitude during interface experiments plotted
against expected amplitude from (b). Error bars represent range of values.
32
dependence of laser displacement against applied galvo voltage, calculated by trapping 500nm
fluorospheres in water. On average the conversion for a ~100nm excursion was 0.0027V and 0.0037V
for Y and X movements respectively (fig 2.2 b). In practice, the measured laser movement showed an
error of 30-40 nm compared to expected but the fluctuation of trap amplitude for any given position
calibration was 2-4% (~20±10 nm for 0.4-1 µm amplitudes) . (fig 2.2 c, standard deviation too small
to show on graph).
2.4 Optical tweezers experiments
Individual cell-cell interfaces in embryos were mechanically probed in a custom-built light sheet/selective
plane illumination (SPIM) setup coupled with a single beam gradient optical trap. A 100x water-
immersion lens (1.1NA) Nikon was used for imaging the sample on an EMCCD camera (supplementary
figure 1) as well as for introducing the optical trap in the imaging plane. Imaging was done using 488nm
and 561 nm excitation lasers (at 10mW and 60mW laser power). Laser position could be controlled
independently in X and Y with various functions using conjugated galvanometric mirrors (galvos). The
laser position in the camera image was calibrated using 500 nm carboxylate coated flurophores
(Invitrogen, excitation/emission peaks 580/605nm). In practice, the optical trap was sufficient to excite
the beads due to 2 photon absorption. The beads were trapped at ~50-100 mW laser power (after
objective) and given a circular oscillation of 10 µm at 0.07 Hz. All these parameters were standardized
after trial and error by checking the calibration accuracy with the beads. Nevertheless the calibration can
vary for larger excursions after just a few hours, therefore it was done for every experiment when
possible). The entire protocol for manipulating and recording the experiment was developed with custom
scripts and GUIs (Claire Chardes, Olivier Blanc) in Qt Creator (C++), and the triggering as well recording
of the voltage sent to the galvos was synchronized with the trigger for the camera recording (which could
be done in the range of 8-30 fps, depending on the experiment). The recordings were then analysed in
Matlab, using custom scripts, to calibrate the conversion of galvo voltage to trap position on the camera
image. For manipulating the cell-cell interface, initially, sinusoidal oscillations were given in X, Y or X-Y
axis, with different time periods ranging from 0.3- 5 seconds and amplitudes from 0.3-1.1 µm. Laser
power was varied from ~50-200 mW (after objective). These experiments had to be done during the end
of cellularization till the beginning of gastrulation (to ensure minimum variation in tissue properties).
Cycles recorded depended on period frequency but ranged from 2-20. After measuring the effect of these
parameters on deflection measurements, values were kept constant with time period of 2 seconds, laser
amplitude ~0.5 µm and laser power ~200 mW. Experiments were done in the cadherin plane (1-3 µm
33
from the apical cortex, depending on embryo stage) as well as in a plane that was 3 µm basal; at the end
of cellularization as well as during early germband elongation. In addition, pull-release experiments were
also done, where the laser was kept stationary and switched on at a distance of a few 100 nm-2 µm from
the cell-cell interface for 10s-1minute and the interface deformation as well as subsequent relaxation were
recorded.
2.5 Data Analysis
Kymographs of interface deflections were initially produced from the camera movies manually in Fiji
(multiple kymograph plugin, line thickness 6-9 pixels, with the resolution of the setup being 194 nm per
pixel). Using a gaussian fitting super-localization matlab script (developed earlier in the lab by Olivier
Blanc) on the kymograph, the cell-cell interface could be localized with sub-pixel accuracy. From this, a
minimum of 2 to a maximum of 10 oscillations (for fast frequencies >2hz) were used for measuring the
amplitude of deflection. However, the laser position calibration, along with additional matlab scripts
developed recently (Claire Chardes, Raphael Clement), can obtain a semi-automatic kymograph
(autokymo) and a sub-pixel localization of both laser and interface position directly from the recorded
images and voltage. In addition to measuring amplitude, this provides information on the spatiotemporal
relationship of laser-interface interaction. This position displacement data was therefore used to measure
the phase lag and position offset between the laser-interface. Data was stored in either Microsoft Excel
sheets or matlab figure and data files. Graphs were obtained using either Excel or matlab. Statistics was
done using the unpaired T-test (Graphpad).
2.6 Experimental limitations
There were variations in membrane intensity resulting from aberrations inherent in light sheet
microscopy (scattering and shadow effects) and 2 photon-excitation/bleaching that was observed in every
interface (fig 2.3 a) (excitation peak for cherry is close to 561nm which can be excited by the 1070nm
infra red trap). The excitation laser power was also varied to account for different acquisitions speeds at
different time periods of amplitude. Hence exact localization accuracy for the trap experiments could not
be calculated but amplitude of noise for stationary interface without any trap was consistent at different
stages between ~30±10 nm (fig 2.3 b). In addition the autokymo always optimizes a kymograph
34
perpendicular to laser movement axis, however in many cases this was not the best axis for measuring
interface deflection, due to stochastic variation in interface orientation as well as the rapidness with which
Figure 2.3 Noise in interface position detection. a) Interface deflection (black) and interface intensity
(green, AU) plotted as a function of time. Images correspond to interface image at the respective time
points (red=trap position). B) Amplitude of stationary interface noise (no trap) at different embryo stages.
Error bars are standard deviation.
the embryo could develop. Experiments in germband elongation were furthermore affected by tissue drift
and thus measurements were limited to 2-3 cycles and fast time period (2s or less). Also when measuring
offset between interface and laser positions, there is an uncertainty in laser position resulting from input
voltage noise (20±10 nm) and localization error of beads used for calibration (20-30 nm). Localization
accuracy for trap-interface offset could therefore be estimated to be ~ 70-100 nm. Indeed the offset of
expected to calibrated laser amplitude is also 20-40 nm (fig 2.2 c).
An obvious concern with a high power infra-red laser, is thermal damage to the tissue. While we did not
calculate the point spread function (PSF) of the trap, an objective NA of 1.1 and laser wavelength of
1070nm gives a PSF estimate of ~500nm (lambda/2NA). It has been shown under trapping conditions
35
similar to my experiments, that an instantaneous temperature increase of few degrees can be induced
locally (few 100 nm) on laser exposure, which decays within a second when the trap is switched off
(Ebbesen et al. 2012),. In my experiments, the major concern with this temperature increase is
of thermally induced breakage/perturbation of the cell-cell interface. Indeed at 400mW laser powers
we do see some damage (in terms of irreversible calcium uptake and interface shape changes). We do
not see this damage in 200mW power used for most experiments. Additionally, the interface
movement amplitude does not increase over several cycles in control embryos and the intial position of
the interface is restored after removing the trap in most cases. But to be on the safe side most
measurments were done for less than 30s per interface and only the first few cycles/oscillations were used
for analysis.
2.7 Quantification of E-cadherin, myosin II and lifeact
Imaging was done in a Perkin-Elmer spinning disk microscope using a 100x oil immersion lens. Z-stacks
from 6-12 µm thickness were acquired, starting from the apical cortex, with a slice thickness of 1 µm and
a sampling rate of 2 Hz (along with low laser power of 20% to minimize bleaching). Single channel
imaging was done with either 488nm excitation alone or sometimes (especially for myosin-II and ROCK
inhibitor where tissue could become largely invisible) squash-Gap43::cherry images were acquired with
the 561 laser for reference. Analysis of images was done with Fiji and the Perkin Elmer plugin. Maximum
intensity projections of 3 slices (with the plane of interest centered) were taken and then background
subtracted (ImageJ, 20 pixel rolling ball radius for E-cadherin and myosin-II, 60 pixel for lifeact). The
mean junctional intensity (AU-arbitrary units) of individual junctions in the resulting images was then
calculated and averaged.
36
CHAPTER 3
ESTABLISHING OPTICAL TWEEZERS
AS AN IN VIVO TOOL TO UNDERSTAND
CELL-CELL INTERFACE MECHANICS
37
38
3.1 Motivation
Optical tweezers and microrheology remain the most promising techniques in vivo due to their non-
invasive nature and precision. Previous work in our lab was directed towards developing and applying
these techniques to measure directly the forces involved in germband elongation. But these techniques
remain limited by the requirement of external probes (typically beads). One way to overcome this
limitation is to use natural probes like lipid granules or vesicles. Since germband elongation is
immediately preceded by cellularization, it is possible to inject classical microsphere probes at the end of
cellularization and study their dynamics during gastrulation. This however limits the success of the
measurements to the few cell-cell interfaces where the beads successfully localize without interacting
with the cellular cortex. I tried to overcome this limitation by developing a new approach to probe cell-
cell interfaces directly in Drosophila embryos.
39
Figure 3.1 Characterizing the deflection of cell-cell interfaces imposed by optical tweezers
(A) Schematic of the setup: the embryo is optically sectioned by a light sheet and imaged, while a laser
trap (red) allows manipulation. Top image shows the epithelium labeled by a membrane marker
(GAP43::mcherry) and the laser trap position (marked by a yellow arrowhead). (B) Schematic of
deflection with distribution of forces and overlaid as well as separate images of the interface in mean and
the extreme positions of deflection. Scale bar: 5µm. (C) Representative plot of deflection versus time
showing both trap (red solid line) and interface positions (black solid line). (D) Interface deflection as a
function of laser trap power. (E) Interface deflection plotted against junction size (stage 6 end)
40
3.2 Results
3.2.1 Characterizing the deflection of cell-cell interfaces imposed by
optical tweezers
The mechanics of cell-cell interfaces was probed using a setup combining optical tweezers and light sheet
microscopy (Fig. 3.1A). This setup allows imaging of the embryonic epithelium at a high acquisition
rates, while manipulating objects in vivo. The laser trap was custom built by me and is produced by a
near-infrared laser light focused by the collection objective lens into the sample and is moved by
galvanometric mirrors in the plane of the epithelium (Fig. 3.1A and Fig. 2.1). Optical tweezers
experiments usually require a probe (glass or a polystyrene bead) to apply forces on attached molecules or
structures. This can be problematic for in vivo experiments which would require the probes to be
injected externally. I found that the cell-cell interfaces can be manipulated directly, without the need of an
external probe (Fig. 3.1B. Inset showing 3 snapshots of deflected membrane interfaces and trap
position.). Applying the laser directly, I imposed a sinusoidal movement to the trap perpendicular to a
cell interface and centered on it, and imaged the resulting deflection in the adherens junction plane
(Fig. 3.1C). The interface deflection followed the trap movement with lower amplitude, indicating that
the interface resists the mechanical load imposed by the laser trap (Fig. 1C).
The laser power was then varied while keeping trap sinusoidal amplitude �� (0.5µm) and the period of
oscillation (2s) constant. I found that �� also increases linearly with the laser power up to 300 mW (Fig.
3.1D). To explore the regime of deformation that the laser trap imposes to the interfaces, I then varied��, while keeping the period of oscillation constant (2s). The amplitude of the interface deflection ��, increases with ��, yet it deviates from a linear relationship for trap amplitude larger than 1 µm (Fig. 3.2b).
The results were similar even when the velocity was kept constant (2µm/s) by increasing the time period
with amplitude (Fig. 3.2a). Thus the variation is produced largely by the amplitude if the velocity is slow.
Together, these results imply that in the regime of small deformations (for an average interface length of
4.5 µm), the trap acts as a linear spring, whose stiffness �� is linearly proportional to the laser power.
Furthermore, �� does not vary with interface length at constant ��, time period and power (Fig. 3.1E).
Hence all further experiments were carried out within this range of deformation (<1 µm trap amplitude)
and with power 200 mW, unless otherwise stated.
41
Figure 3.2 Interface deflection as a function of trap movement amplitude at, a) constant velocity (2µm/s)
and b) constant time period (2s). Interface deflection (c) Interface deflection as a function of the trap
oscillation period. (amplitude: 0.5µm, period: 2s). (fitting in b) and c) by Raphael Clement)
42
The resistance to deformation can arise not only from the mechanical properties of the interface and its
cortical elements, including the acto-myosin cytoskeleton, but also from the viscous drag force exerted by
the cytosol. To determine whether the resistance to deformation is time-dependent, we varied the period
of oscillation, while keeping the trap amplitude constant (amplitude At = 0.5 µm, Fig.3.2c). For periods
larger than or equal to 1 s, which correspond to mean speeds smaller than 2 µm.s-1 the amplitude was
constant. We observed that for periods smaller than 1 s (speed > 2 µm.s-1), the deflection was reduced
(Fig. 3.2c), which is the characteristic signature of viscous damping, presumably related to the viscous
drag in the cytosol.
43
Figure 3.3 Offset between laser and interface positions
a) Schematic of trap-interface displacement curve, defining initial offset (ini_off), final offset (fin_off)
and interface movement (int_mov)
b) Final offset plotted as a function of initial offset
Interface deflection amplitude as a function of c) final position offset (data from 1,2,4 and 5 second
periods for 0.5 µm trap amplitude were pooled together, red: linear fit, blue: data points for individual
measurements),and d) initial offset.
(for c and d, measurements with initial offset greater than 600 nm were rejected. Negative values of final
offset in b) represent movement away from the laser. For c), all values were taken as positive)
ini_off
fin_off
int_mov
44
3.2.2 Dynamics of cell-cell interface interaction with optical tweezers as
a measure of position offset and phase lag
While the trap is positioned on the junction during the experiments there can be an uncertainty of few
100nm (pixel size of camera is 194nm). Therefore it was important to track the position of the trap with
respect to the interface during the experiments. It was observed that the interface often does not oscillate
at its rest position but instead acquires a new position to oscillate (Fig. 3.3a). Therefore I defined initial
(ini_off) and final offset (fin_off), as the distance between the trap and interface mean positions before
and during the experiments (Fig. 3.3a). Addition of ini_off and fin_off gives the interface movement
(int_mov) during the experiment. Fin_off varies linearly with ini_off (Fig. 3.3b). For small initial offset
values (<600nm) the final offset remained between 0-500nm. Importantly the final offset tended to be
positive in most cases indicating tendency of the interface to move towards the trap (negative final offset
indicates movement away from the trap, in the example in 3.3a, fin_off is positive). Assuming a trap size
and cell-cell interface thickness of few 100nm, this implies that the interface is always under the influence
of the trap during the experiments. Nevertheless final offset can have a small but measurable influence on
interface deflection amplitude �� (Fig. 3.3c). As fin_off increases, �� decreases linearly. For final offset
values within 500nm we see a variation in �� of ~500nm. It is possible that this variation might arise to
some extent due to variation in interface tension. Nevertheless we do not see any correlation between
ini_off and ��(Fig. 3.3d).
Since we observed a weak time dependence of �� it was interesting to look at the phase lag between trap
and interface oscillation. A purely elastic behavior would show 0 phase lag while a purely viscous
behavior can show a phase lag of 90°. Between 2-5 seconds of time period of oscillation, phase lag
remains consistently between ~5-10° (Fig. 3.4 a) implying minimal viscous drag. Experiments below 1
second were hampered by noise in the interface oscillation. But at 1 second of time period we do see a
larger scatter in phase lag (Fig. 3.4a). We then looked at the variation of phase lag with int_mov (Fig.
3.4b) and final offset (Fig. 3.4c). Interestingly, phase lag increases linearly with int_mov but seems to
decrease with increasing final offset for small values of fin_off (<300nm).
45
Figure 3.4 Phase lag between laser and interface movement as a function of a) time period (a), b)
interface movement (Initial offset – final offset) and c) final offset. (The mean trap position is defined as
the zero axis. For b and c, data with net movement <100nm and/or final offset larger than 700nm were not
taken into consideration. All values were taken as positive for plotting) Error bars are standard deviation.
46
3.2.3 Relaxation of interface deformation
We performed pull-release experiments, which have been used in vitro on single cells with optical and
magnetic tweezers (Alenghat et al. 2000) but have never been applied in vivo. Pull-release
experiments consist of switching the laser trap on/off at a few hundred nanometers distance from the
junction and then monitoring the deflection of the cell-cell interface, both towards (trap on) and away
from the trap (trap off) (Fig. 3.5a). The pull-release curves obtained show that the dynamics are not
purely exponential, and exhibit at least two characteristic times, in the range of 1 s and 10 s (Fig. 3.5a).
3.2.4 Propagation of interface deformation
A challenging question in tissue morphogenesis is whether local forces produce long-range deformation,
and at what speed the mechanical information propagates. We imposed the local deflection of a cell
interface using sinusoidal oscillations, to see how single cell deformation propagates throughout the
tissue. We tracked the deflection of neighbouring interfaces away from this point. For that purpose, we
plotted kymographs along lines perpendicular to cell interfaces (Fig. 3.5b). We oscillated the target
interface using a deflection amplitude of 1 ± 0.1 µm (Fig. 3.5b, bottom panel). The neighbouring
interfaces within a distance of 1 to 2 cells also deflected periodically, but with much lower amplitudes and
a small phase shift (Fig. 3.5e), top panel). This indicates that the deformation typically decays over a
distance of 1-2 cells.
Having observed the local propagation of interface deformation, we then wondered if we could deform
the interface without trapping it directly. For this purpose we introduced the trap in the cytoplasm of the
cell (Fig. 3.6a) and oscillated it at a 2 second period. With a cell size of ~6-7 µm, it is quite likely that one
junction would be closer to the trap than the others. Moreover we used a trap displacement of 1-3 µm. In
these conditions, the junction which deformed the most was defined as i (assuming it was closest to the
trap). The displacement of this junction was used as reference to also track ii, iii and iv. All the junctions
oscillate in response to the trap movement with an amplitude range of 50-200nm although the amplitude
of iv is significantly reduced (Fig. 3.6c). As the trap amplitude is increased, the interface amplitude
increases reaching a value of 50% of a normally trapped interface at 1.5 µm (Fig. 3.6c). Interestingly the
interface movement is much larger compared to deflection amplitude (Fig. 2b, negative is movement
away from trap) but shows no net movement towards the trap on average. We also measured the phase lag
of junctions ii and iii in comparison to i (Fig. 3.6d). There is a significant phase lag at 0.5 µm trap
amplitude (20-30°) which decreases linearly with increase in amplitude.
47
Figure 3.5 Mechanical model of the interface and tissue response (modified from Bambardekar et al. in
press. Analysis, schematic and model by Raphael Clement) a) Deflection of the interface in a pull-release
(trap on - trap off) experiment. The model (blue line) fits accurately the experimental data (black). Notably,
the relaxation is not a simple exponential. The simplest analogous visco-elastic model (inset): a maxwell arm in parallel with a spring, moving in a fluid of viscosity η (the cytosol). b) Deflection perpendicular to
the interfaces is tracked in time along lines perpendicular to cell interfaces (red), as a kymograph, which
allows measuring deformation away from the targeted interface. c) A viscosity decrease in the simulations
results in increased propagation, whereas a viscosity increase results in decreased propagation. d) Overlay
of the rest (purple) and deformed (green) tissue in the experimental (left) and simulated (right) tissue. e)
Spatial decay of interfaces deflection over the neighbouring cells. The model quantitatively predicts the
decay.
48
Figure 3.6 Interface deflection induced by cytoplasmic trap. a) Schematic of experiment. Cell-cell
interfaces in violet. Trap (red) is oscillated in the apical cytoplasm (black arrows). The junctions
perpendicular to the trap movement get drawn towards the trap. The junction that deforms the most is
taken to be I as reference for identifying neighbouring junctions. b) Interface deflection amplitude at
different amplitudes of trap excursion in the cytoplasm. c) Interface movement at different trap
amplitudes. d) Phase lag between junctions (i and ii/iii) at different trap amplitudes. Error bars are
standard deviation.
49
3.3 Discussions
3.3.1 Origin and nature of in vivo optical forces
Figure 3.7 (A) Quantitative Phase Microscopy Image obtained on early Drosophila embryo (Stage 6).
The method uses a transmission light microscope and quadriwave lateral shearing interferometry as
described in (Bon et al. 2009) . The epithelial cells are observed in cross-section (cartoon, top) Scale bar, 5
µm (white). The calibration bar shows the optical path difference in nm. (B) A plot profile along a line
(red line in (A)) shows that the optical path difference is larger at cell interfaces than inside the cells.
Given that the line defines positions where the geometrical thickness of the embryo is constant, this
indicates a refraction index increase at cell interfaces. Arrows of different colors mark the position of 3
interfaces.
Optical trapping results from a positive refractive index mismatch between the trapped object and the
surrounding medium (Ashkin 1997). Indeed, trapping of particles within the cytoplasm has
been previously reported (Ashkin 1989, Welte et al. 1998). Furthermore the lipid membrane can be
trapped and manipulated directly, however the forces produced are very weak. Thus the membrane itself
is unlikely to be responsible for the deformation observed in the interface. To check if the interface
has a positive refractive index compared to the cytoplasm, we performed quantitative phase imaging
of the epithelial cells (Fig.3.7). Indeed there is a positive mismatch of refractive index. Yet, the value of
this mismatch is difficult to determine because of the geometry of the tissue. Therefore, to estimate the
trap stiffness on the interfaces, the deformation produced by direct application of the focused laser was
compared with that induced by 0.46 µm diameter beads pushed against the cell-cell interfaces (Claire
Chardes, Fig. 3.8). The former was only by 2- to 3-fold larger than the latter (2.5 ± 0.4, mean±SD, 4
measurements), indicating that the trap stiffness on the interfaces was 2- to 3-fold smaller than that on50
beads. The trap stiffness on beads was 120 ± 50 pN.µm-1 at 200 mW laser excitation (20 measurements),
thus the trap stiffness on interfaces was estimated to be 50 ± 30 pN.µm-1, in the regime of small
deformations.
Figure 3.8 Interface deformation with and without beads. (A) Snapshots of an interface deformation
induced by a 500 nm diameter bead, moved by the laser trap. The red and green channels correspond to
two different positions of the trap separated by approximately 0.5 µm. (B) Positions of laser trap, bead
and interface in an oscillatory experiment at 100 mW laser power. (C) Positions of laser trap and interface
in the same conditions as in (B) in the absence of bead. (Claire Chardes)
Other than sinusoidal amplitude, another way to understand the optical forces produced in our
experiments, is the interface movement. Given that the trap-interface distance tends to have a positive
final offset is a clear indication that the forces produced are attractive (Fig. 3.3a,b). If such was the case
the interface should deform less with increase final offset during the oscillations, whereas the initial offset
should have no bearing by itself on interface amplitude. Indeed this seems to be the case (Fig. 3.3b and c).
Interestingly, placing the trap ~3 µm away from the interface still produced interface movement as well as
oscillation which also seems to propagate to the neighbouring junctions (Fig. 3.6b and c). This oscillatory
movement is only 2-4 fold smaller than when the interface is trapped directly (trap-interface distance 0-
600nm). For small deformations (trap excursion 1 µm) this might indicate a linear relationship between
interface amplitude and final offset (for a range of 0-3 µm), implying that the force produced is linear
51
akin to a spring. Furthermore, the propagation in this case is due to optical attraction rather than tissue
mechanics, as the interface further away from the trap but at a similar distance from interface i compared
to interface ii and iii shows a very weak oscillation at all excursions (Fig. 3.6c). Additionally the phase
lag between the interfaces decreases with increase in trap excursion (Fig. 3.6d) which one would expect in
this scenario since the trap varies its proximity to interfaces i and ii at different time points in the
oscillation (Fig. 3.6a). Large fluctuations in final offset for all interfaces (Fig. 3.6b) probably imply
stochastic mechanical changes produced in the cell by the cytoplasmic trap and need to be further
investigated to be understood properly.
Figure 3.9 Viscosity measurements obtained from trajectories of individual beads. (A) Image showing
100 nm diameter beads (red) injected in the embryo. Cell contours are labeled by E-cadherin::GFP. (B)
Single particle trajectories superimposed to an image of the cells. (C) Fraction of beads exhibiting
diffusive, subdiffusive and superdiffusive behaviours from analysis of the mean-square displacement
(using criteria as described in (32)). Trajectories were acquired at 38 Hz over a time from 13 to 26 s. (D)
Histogram of the viscosity coefficient deduced from the analysis of 1348 particles, exhibiting free-like
diffusion. (Olivier Blanc)
52
3.3.2 Towards a mechanistic understanding of the experiments
The constancy of the deflection amplitude at speeds below 2 µm/s indicates that the force applied by the
trap and that produced by the interface are in quasi-static equilibrium. Interestingly, this condition of low
speeds (< 2 µm/s) is always fulfilled during cell shape changes driven by Myo-II contractility during
tissue morphogenesis of Drosophila, indicating that short-time viscous damping should play no role in the
process.
In the quasi-static regime (deformation speeds < 2 µm/s), we can thus assume that the shape of the
interface mainly results from the balance of forces between the trapping force �� and the tension of the
interface �: �� = �� ��s �., where � is the angle that the interface makes with respect to the trapping
force (Fig. 3.1B). As the vertices of the cell-cell contact did not move significantly during the
deformation (Fig.3.11), we could neglect the contribution of other cells and use this simple local
equilibrium formula.
For small deformations (that is for maximal deflections much smaller than the initial junction length
l0), ��� � ≈ ���� and �� ≈ ��(�� − �), where � is the position of the interface and �� the position of the
trap. The tension of the interface thus approximates as: � ≈ ����� ���� − ��. We found that the ratio ��� remains almost constant during periodic oscillations (Fig. 3.2c), indicating that the pre-existing tension
of the interface is not significantly modified during small deformations. Importantly, this also implies that
tension measurements, while relying on geometrical and physical approximations, do not require a
mechanical model of cell contacts. Thus, tension values can be obtained as a simple linear function of the
ratio between the interface deflection and the trap position. From the estimated trap stiffness on the order
of 50 ± 30 pN.µm-1 (Fig. 3.8), and using the slope in Fig. 3.2c as a mean value for the ratio, tension at
cell-cell interfaces � is estimated on the order of 60 ± 40 pN. Given that a single molecular motor of
Myo-II at maximal load produces 5 pN of force, the range of tension measured here suggests that a net
imbalance of about a dozen of motors pulling on a cell interfaces could deflect it by a few hundred
nanometers. The tensions reported here are 2-3 orders of magnitude below cell-cell forces in cell
aggregates on adhesive substrates in vitro (Maruthamuthu et al. 2011)
We do see a measurable viscous damping effect at oscillation periods smaller than 1s (Fig. 3.2c). This is
further indicated from the increase in phase lag range at 1s (Fig. 3.4a, although measurements at smaller
periods could not be done). The increase in phase lag with interface movement (Fig. 3.4b) might be an
53
implication that the damping arises from the cytosol. To test whether damping was due to the viscosity of
the cytosol, we performed relaxation experiments after release of the trap, which provide an
Figure 3.10 Relaxation of interfaces after trap release. Initial speed of relaxation is related to the
friction coefficient. (Raphael Clement)
alternative measurement of tension. At the onset of relaxation, tension is balanced only by viscous
damping: �� ��� � = ����, where v0 is the initial relaxation velocity and �� is the friction coefficient.
Measuring v0 (Fig. 3.10) and using the mean tension value of 60 pN, this provided an indirect
measurement of ��, which is in the order of 1.2 ± 0.6 10-4 m.Pa.s. To determine viscosity, �� should be
rescaled by both a typical length scale (the junction length, 4.5 µm) and a geometric coefficient (=16 for a
circular disk, which we will use here to obtain an order of magnitude). The viscosity associated with the
observed damping is thus in the order of a few Pa.s. To confront this estimate to the actual viscosity of the
cytosol, beads of radius � = �� �� were injected in the cytosol (He B et al. 2014) and their mean square
displacement was measured (Olivier Blanc, Fig. 3.9). Relating the diffusion constant � to the viscosity �
by the Stokes-Einstein equation (��� = �����), where ��is the Boltzmann constant and � the
temperature, the viscosity of the cytosol is found to be 3.6 ± 0.1 Pa.s (mean ± SE, 1350 beads). This is
consistent with the order of magnitude found with the relaxation method, which confirms that damping
might indeed be caused by the viscous drag in the cytosol.
54
Figure 3.11 Elongation of interfaces adjacent to the optically deformed interface. (Raphael Clement)
3.3.3 Developing a mechanical model (Raphael Clement)
Different types of visco-elastic models were considered to fit the experimental data and it was found that a
so-called standard linear solid (SLS) model, composed of a Maxwell arm (a spring and a dashpot) in
parallel with a spring and embedded in a viscous medium of viscosity � (associated with the cytosol), is
the best and simplest model to correctly account for the observed behavior (Fig. 3.5a; fit: solid blue curve,
model: inset). Importantly, this simple model also fits well the experiments using a sinusoidal movement
described above (Fig. 3.2b & c, solid blue curve). At these short time scales, the mechanical response of
an epithelial tissue should result both from the constitutive mechanics of its acto-myosin cortex and from
the viscosity of the cytosol. The timescale of 1 s is consistent with the drop of deflection amplitude
observed below 1 s in the oscillatory experiments (Fig. 3.2c), and can be attributed to the viscous drag in
the cytosol. Consequently, the existence of another time scale (10 s) suggests that the cortex itself is not
purely elastic but visco-elastic (hence the SLS model). Importantly, the visco-elastic dynamics
characterized here are on a short time scale (under one minute). Neither the experiments nor the model
concern the long-term dynamics (minutes to hours), which presumably implies creep and therefore fluid-
like behavior (He B et al. 2014).
55
Single junction
The mechanical model for a single junction is derived from the constitutive equation of the standard linear
solid model and from a force balance equation at the interface. The constitutive equation relates the
horizontal force � pulling back the interface, to the horizontal displacement � of the interface:
�̇ +��� � = (�� + ��) �̇ +
��+��� �,
where �� and �� are elastic parameters (N/m), � a viscous parameter (m.Pa.s) for the cortex and the dot
denotes a temporal derivative. At these very low Reynolds number, inertia can be neglected, and the
balance of forces at the interface then simply reads: � = ��( �� − �) − ���̇,
where �� and �� are the stiffness and position of the optical trap, respectively, and �� is the friction
coefficient of the interface in the cytosol. The first term on the right-hand side thus corresponds to the
force exerted by the optical trap, while the second corresponds to the viscous drag in the cytosol – and is
therefore proportional to the velocity �̇. This linear system can then be solved for any trap trajectory ��(�); in particular for our experimental conditions: a sinusoidal oscillation or a pull-release experiment.
Tissue scale
In the tissue scale simulations, the epithelium is considered as a network of bonds – the cell contacts –
between vertices. Each bond is considered as a visco-elastic segment. The constitutive equation of each
segment, similar to the first equation, is:
�̇ +�2� � = (�1 + �2) �̇ +
�1+�2� �,
where � is the tension increase and � is the length increase, � = � − �0. The motion of each vertex is then
computed using the force balance equation (Fig. 3.11). For the vertex i:
�������⃗̇ = � �������⃗�=���(�)
56
Notably, the midpoint of the target interface is treated as a vertex in the simulations. Its position is
imposed as to mimic the considered experiment. The rest of the vertices move according to the force
balance equation, therefore, their movement ultimately results from the deflection movement of the target
interface.
To evaluate the ability of this simple mechanical model to reproduce these data, the single junction
model was first transposed to a network of contacts, as observed in vivo. Each contact is
considered as a viscoelastic (SLS) element pulling on vertices, so each vertex motion results from the
balance between tensile forces and fluid friction (Fig. 3.5d). It was then possible to simulate the
mechanical response of the tissue to periodic manipulations in the extracted experimental geometry,
imposing only the sinusoidal displacement at the midpoint of the target interface. The kymographs in the
simulated tissues were plotted along the same lines as in the experiments (Fig. 3.5c, bottom panel).
The typical parameter values that was estimated from the single interface experiments faithfully
reproduced the tissue-scale observation: one or two neighbouring cells deform away from the source
point of deflection (Fig. 3.5). In contrast, taking higher or lower viscosity for the cytosol into
account leads to shorter and longer propagation distance, respectively, which does not correctly
reproduce the observed behavior (Fig. 3.5). The speed of propagation can be experimentally
estimated by the phase delay between the trapped interface’s deflection and that of its neighbors.
At a 1-cell distance (~7 µm), a time delay of 375 ± 125 ms was measured, which corresponds to a
propagation speed with a typical phase velocity of 20 µm/s. At a 2-cell distance, a mechanical
signal would propagate in less than 1 s. This is much faster than any diffusible chemical; for instance,
a protein diffusing at 1 µm2.s
-1 would explore the same distance in 100 s. This speed is also much
larger than the speed of acto-myosin flows, which are observed in different systems, including
Drosophila and C. elegans (0.1 µm/s) (Mayer 2010).
57
3.4 Summary
To summarize, cell-cell interfaces in Drosophila embryo can be directly manipulated with a laser trap.
This trapping arises from a positive refractive index mismatch between the interface and the cytosol. The
temporal and spatial kinetics of this manipulation were analyzed by sinusoidal excursions and pull-release
experiments. For small excursions and final offsets the force produced is linear and quasistatic. Both
manipulations reveal the role of viscous damping at time scales below 1 second.
Interestingly the sinusoidal excursion can propagate to neighbouring junctions. The propagation decays
within a distance of one cell-length and shows only a small phase lag. Cytoplasmic trap experiments
showed very different behaviour by comparison, indicating that the propagation seen after trapping
interface was mechanical in origin.
Using my experimental data a passive viscoelastic model was developed which successfully replicates the
findings from the experiments. Having standardized the method for a passive tissue (apical epithelium at
the end of cellularization), this sets the stage to understand mechanical forces in epithelial rearrangements
during gastrulation.
58
CHAPTER 4
PROBING MECHANICAL ANISOTROPY
DURING DROSOPHILA GERMBAND
ELONGATION
59
60
4.1 Motivation
Having established methodology and a viscoelastic model for passive epithelium (Chapter 3) we decided
to probe epithelial mechanics in early gastrulation of the Drosophila embryo. The early epithelium
consists of a simple sheet of cells that spread over the yolk and are in contact with each other through E-
cadherin-based adhesion. During early embryogenesis at the blastula stage just after the end of
cellularization, epithelial cells have very similar hexagonal shapes, suggesting that cell junctions have
similar mechanical properties and the internal pressure of these cells is homogeneous. At the later gastrula
stage (germband elongation), cells undergo shape changes at distinct regions in the embryo. On the
ventral side of the embryo, apical cell constriction of a few rows of cells drives tissue invagination
(Kolsch et al. 2007), while on the ventro-lateral side of the embryo, cell intercalation, a process
whereby cells exchange neighbors by polarized remodeling of their junctions, drives tissue
extension. Laser dissection of cortical acto-myosin networks at cell junctions in the ventro-lateral
tissue has shown that anisotropic distribution of Myo-II causes an anisotropic cortical tension (Rauzi et
al. 2008). However, the absolute values of tensile forces have not yet been measured, and more
generally, the mechanics of cell-cell interfaces in vivo is largely unknown. These questions are addressed
in the following section by local mechanical measurements at cell junctions during germband elongation.
61
Figure 4.1 Tension at cell contacts before and during germband elongation in wild type embryos.
a) Schematic of tissue elongation and images at three different stages: before elongation (stage 5 end), at
the onset of Myosin-II accumulation at cell junctions (stage 6 end) and during tissue extension (stage 7)
showing the cell interface (pink, GAP43::mcherry), E-cadherin (green) and Myosin-II (red). b) Interface
deflection amplitude at the adherens junction plane, for different stages t-Test (* p-value<0.05 ) c)
interface deflectionfor different stages normalized to junction size. d) apical E-cadherin levels
(endocad::GFP) in different stages. e) apical myosin-II levels in different stages. t-test ( *** P<0.001)
Scale bar 10 µm. Error bars are standard deviation.
62
4.2 Results
4.2.1 Probing the anisotropy of mechanical forces in early gastrulation
Interface deflection amplitude was measured at different stages of Drosophila embryo. At stage end the
epithelium is finishing cellularization. At stage 6 end mesoderm invagination ends and germband
extension begins. Stage 7 is early germband extension (Fig. 4.1a). The amplitude of deflection of the
cell-cell interface is reduced by two-fold from the end of stage 5 to the onset of stage 7 (Fig. 4.1b). The
decrease in amplitude indicates an increase in tension which is accompanied by significant enrichment of
E-cadherin and Myosin-II (Fig. 4.1d and e). Comparing the tensions for different orientations at this stage
revealed that the cell-cell interfaces along the D/V axis with myosin enrichment (Fig.4.2d) are about 2.5
times more tensed than those along the horizontal axis which are enriched in cadherin and have less
myosin (Fig. 4.2 c and d), which corresponds to mean absolute tensions of 275 ± 137 pN and 108 ± 54
pN, respectively (Fig. 4.2a). We probed the tension of interfaces of lengths varying from 4 to 8 µm, but
normalizing the deflection amplitude to junction length leaves our conclusion unchanged (Fig. 4.1c, Fig.
4.2b).
Measurements of interface deflection by laser trap at different positions along the apico-basal axis showed
that during tissue morphogenesis, there is a gradual polarization of the tension along this axis (Fig 4.3d)
concomitant with development of apico-basal polarity of cadherin and myosin (Fig. 4.3a,b and c). By
comparison apical and basal deflection remains unchanged in stage 5 end. To see if cadherin clusters can
regulate tension in the absence of myosin we probed stage 5 end further in the lateral plane (Fig. 4.4a).
Cadherin is localized in the first ~0-7 µm at this stage (Fig. 4.4a). Hence all measurements in this range
were taken as apical. Basal measurements were made between 8-20 µm. Myosin activity is not seen in
this distance (at the end of cellularization myosin II was localized basally at the cellularization front, as
seen in SPIM as well as spinning disk imaging). We find that the interface deforms more basally
indicating a reduced tension (Fig. 4.4b).
Inhibition of Myo-II activity by injection of ROCK inhibitor resulted in a significant reduction of the
tension at cell junctions (stage 7, tensions reduced to 60 ± 40 pN) (Fig. 4.5c) concomitant with loss of
myosin (Fig.4.5a and b). By comparison, tension remains unchanged after ROCK injection in stage 5 end
(Fig. 4.5d). The value of tension in ROCK injected embryos at stage 7 is similar to tensions measured at
stage 5 end. This confirms that the typical 2-fold increase in tension measured between stage 5 (~60 pN)
and stage 7 (~108 (A/P) to 275 pN (D/V)) can be attributed to myosin-II activity.
63
Figure 4.2 a) Interface deflection amplitude plotted at different stages for interface orientations along
the D-V axis. b) Interface deflection of junctions with different orientations normalized to junction size.
c) Quantification of apical E-cadherin intensity (endocad::GFP) and d) apical myosin-II intensity
(squash::GFP) as function of interface orientation in different stages. t-test (** P<0.01, *** P<0.001).
Error bars are standard deviation.
64
Figure 4.3 a) Images of E-cadherin (endocad::GFP, green) and myosin-II (squash::GFP, red) in the
adherens junction plane and in a basal plane (3 µm below) at stage 7 b) Ratio of basal to apical interface
deflection amplitude at different stages c) Quantification of apico-basal E-cadherin intensity
(endocad::GFP) and e) apico-basal myosin-II intensity (squash::GFP) in different stages. t-test (* P<0.05,
*** P<0.001). Error bars are standard deviation. Scale bar, 10 µm.
65
Figure 4.4 a) Images of E-cadherin (endocad::GFP, green) and interface (GAP43::cherry, pink) in the
lateral plane at stage 5 end. b) apical (0-7 µm from edge) and basal (8-20 µm) interface deflection along
the lateral plane in stage 5 end. c) Images of lateral interface at stationary and maximum deflection (red is
laser position). t-test (** P<0.01). Error bars are standard deviation. Scale bar, 10 µm.
66
Figure 4.5 a) Images of E-cadherin (endocad::GFP, green) and interface (GAP43::cherry, red) in
control(WT) and ROCK inhibitor injected (ROCK) embryos in the adherens junction plane at stage 7. b)
Quantification of apical myosin-II intensity (squash::GFP) in WT, water and ROCK embryos at stage 7.
c) Interface deflection amplitude in the adherens plane in control (WT), water-injected (water) and ROCK
inhibitor injected embryos at stage 7. c) Interface deflection at stage 5 end in WT and ROCK embryos. t-
test (*** P<0.001). Error bars are standard deviation. Scale bar, 10 µm.
67
4.2.2 Effect of actin perturbation on interface tension
Since cadherin dynamics in germband elongation are influenced by actin activity (Cavey et al. 2008),
we looked at actin concentration in different stages. We observed that actin is strongly enriched
apically at the beginning of cellularization (stage 5 beginning), followed by a significant decrease at
stage 5 end (Fig.4.6a and b). Actin enriches again at stage 7 but not to the same levels as stage 5 (Fig.
4.6a and b). Comparing the interface deflection reveals a significantly higher tension in stage 5
beginning compared to stage 5 end and 7. To confirm if actin cortex is responsible for this difference
in tension, we perturbed actin levels by injecting embryos with cytochalasin D (which cuts actin
filaments) or latrunculin A (which blocks actin polymerization), both of which perturb the actin cortex.
Wildtype embryos in stage 7 do not show any apicobasal or D-V polarization of actin distribution (Fig.
4.7a and c). Perturbed embryos show a decrease in actin that is much more significant in latrunculin
treatment (Fig. 4.7b) than cytochalasin (Fig. 4.7c). Interestingly, cytochalsin induces a much stronger
depletion basally than apically (Fig. 4.7c). Deflection amplitude increases after cytochalasin
treatment but perplexingly remains unchanged in latrunculin treatment (Fig. 4.7d). Examining the
interface movement between initial and final offset, however, shows significantly larger movements
after actin perturbation (Fig. 4.7e, fig. 4.9d snapshots).
68
Figure 4.6 a) Images of lifeact (lifeact::GFP) in apical plane at stage 7, beginning (stage 5 beg) and
end of cellularization (stage 5 end). b) Quantification of apical lifeact intensity (lifeact::GFP) in different
stages. c) Interface deflection amplitude in the apical plane at different stages. t-test (*** P<0.001, **
P<0.01, * P<0.05). Error bars are standard deviation. Scale bar, 10 µm.
69
Figure 4.7 a) Apico-basal images of lifeact (lifeact::GFP) in control (WT), cytochalasin D (cyto D)
and latrunculin A (lat A, apical only) injected embryos in apical plane at stage 7. b) Quantification of
apical lifeact intensity (lifeact::GFP) in WT and lat A stage 7 embryos. c) Quantification of apico-basal
lifeact intensity in WT and cyto D stage 7 embryos. d) Interface deflection amplitude in the apical plane
of different stage 7 embryos. e) Interface movement (intial-final offset) in WT, cyto D and lat A stage 7
embryos. t-test (*** P<0.001, ** P<0.01, * P<0.05). Error bars are standard deviation. Scale bar, 10 µm.
70
Figure 4.8 Phase lag and offset in different embryos a) Phase lag between trap and interface
movement in stage 5 end, stage 6 end, stage 7, ROCK (stage 5 end and stage 7), cyto D and lat A (stage
7) b) Interface movement (intial-final offset) in WT, cyto D and lat A stage 7 embryos. c) Final offset
between interface and trap position in different embryos. Error bars are standard deviation.
71
4.2.3 Temporal and spatial dynamics: Position offset and phase lag
measurements in different stages and perturbations
To ensure that all experiments were done under similar conditions, we measured the final offset for
different stages (Fig. 4.8c). We find that the offset does not vary significantly in different stages and
treatments (except ROCK injected at stage 5 end). To verify if interface movement in Fig. 4.7e is due to
large initial offset values, we measured interface movement in stage 5 end, stage 6 end and stage 7 (as
well as ROCK injected embryos) and found no difference (Fig. 4.8b). To look at influence of viscous
damping on dynamic mechanical behavior we measured phase lag in different stages (Fig. 4.8a). Phase
lag showed a much larger range in stage 7 while the rest were all within the same range. Since we
observed massive interface movement after cytochalasin treatment, we looked at this further by
successively increasing the deflection period on the same interface, in wildtype and cytochalasin
perturbed junctions stage 7 (Fig. 4.9 a,b,c). Time dependence in wildtype stage 7 embryos is similar
to stage 5 end (Fig. 4.9d and Chapter 3). In wildtype, the interface typically moves towards the trap but
tries to relax to its initial position when the trap is switched off between successively increasing
periods of deflection (Fig. 4.9a). For cytochalasin, we observe 2 distinct behavior, with some
junctions deforming (cyto D large) much larger than others (cyto D small). Cyto D large shows increase
in interface amplitude with increasing periods (Fig. 4.9d). Moreover it shows massive movements of
the interface towards trap position which subsequently fail to recover to the initial position. (Fig. 4.9c).
For cytochalasin, regions of low actin levels and patches/clusters with higher actin levels are observed
(Fig. 4.7a). Cyto D small also shows significant interface movement which does not recover but
exhibits a much smaller interface deflection amplitude (Fig. 4.9b). (Note in 4.9 b and c, initial offset
is >1 µm, at this offset interface in wildtype embryos show very little interface movement and
decreased amplitude as described in Chapter 3).
72
Figure 4.9 Creep and time dependence after cytochalsin D treatment a) Examples of interface
movement with increasing time periods in cyto D large (drastic movement), water and cyto D small
(small interface movement) stage 7 embryos. (blue=junction, red=trap. Images: interface marked with Gap43::cherry at corresponding time points) b) Interface deflection amplitude in stage 7 water (blue) and
cyto D large (black, only interfaces with large difference > 0.6 µm of initial and final offset analyzed)
injected embryos as a function of increasing time period (s) applied to individual interfaces.
73
4.3 Discussions
4.3.1 Interface tension is regulated by myosin II dynamics
Using direct optical manipulation, I probed how cell-cell tensions change during tissue morphogenesis, in
early germband of the Drosophila. Before gastrulation movements, at the end of stage 5, cells form a
regular lattice with isotropic shapes and isotropic distribution of Myo-II (Fig. 4.1a, 4.2d). Later, at the end
of stage 6 and the onset of stage 7, the total concentration of Myo-II increases at adherens junctions, and
its distribution becomes anisotropic, with higher levels along interfaces parallel to the D/V axis (Fig. 4.1a,
4.2d) (Bertet et al. 2004, Rauzi et al. 2008). This anisotropic distribution of Myo-II has been shown to
drive polarized junction shrinkage and cell intercalation The anisotropy of tension in different stages that
is observed is consistent with the estimates of the relative values of tension from laser nanodissection
(Rauzi et al. 2008). My method thus provides a direct measure of tension with low perturbation.
Moreover we establish further the development of tension polarity regulated by cadherin and myosin
polarity (Fig. 4.2). While normalizing with interface length does not change the trend, it enhances the
difference in tension from stage 5 to 7. This is probably because the analysis was restricted to junctions
between ~4-8 μm. In germband shrinking junctions are stiff while elongating junctions are more
deformable. Therefore the increase in tension difference up on normalization implies that our
measurements were somewhat biased towards elongating junctions.
The E-cadherin junctions, where the experiments were performed, begin to localize apically
during gastrulation. At the end of cellularization junctions are spread over a few microns below the
apical cortex (data not shown) but become restricted to a thin 1 µm section and are localized about 1-2
µm below the apical surface during germband (Harris et al. 2004, TruongQuang et al. 2013) (Fig. 4.3a,
left panel, 4.3b). As myosin-II accumulates at the adherens junction plane at the end of stage 6 and
onset of stage 7 (Fig. 4.3a, right, 4.3c), this might translate into different mechanical properties at the
adherens junction plane as compared to more basal positions. Indeed we do observe an apico-basal
polarity of tension. (Fi.g 4.3). Interestingly, apico-basal polarity precedes gastrulation and resulting
myosin enrichment, when I take the apical displacement of cadherin into account (Fig. 4.4). In this
experiment the apical measurements were extended upto the first 7 µm, to take into account the spread
of cadherin clusters at the end of cellularization, whereas basal measurements were at 8-20 µm. Myosin
is only present at the basal cellularization front at this stage (~25-30 µm). This implies that interaction
of E-cadherin clusters with the actin cortex might be sufficient to influence tension across the entire
columnar cell. However the mechanism involved could be different from the apico-basal anisotropy
74
4.3.2 Actin cortex is responsible for interface tension
Experiments at stage 5 beginning, end and stage 7, clearly show that interface tension correlates in
average with actin intensity at the interface (Fig. 4.6a, b and c). This is in accordance with the general
view that that the cell surface tension is controlled by the actin cortex (Bendix et al. 2008). Notably I
did not observe actin polarization in stage 7 (Fig. 4.6 and 4.7), which is in contrast with the previous
results (Blankenship et al. 2006). Since the marker I used (lifeact) is not as reliable as phalloidin (used
by Blankenship et al.) it could simply be an experimental error. Therefore, it is essential to repeat the
experiments, possibly with phalloidin for comparison. It has been shown that actin exists in 2
different pools in germband epithelia in Drosophila, a smaller pool associating with cadherin
junctions, regulating its stability and a larger pool regulating its mobility (Cavey et al. 2008).
Notably the smaller pool is resistant to latrunculin treatment, but its response to cytochalasin in
unknown. It might be crucial to look at the dynamics of these pools during the experiments as an
alternative explanation for the observed discrepancy.
The large scatter in cyto D interface amplitude as well as the smaller amplitude in latrunculin A is
perplexing (Fig. 4.7d). However actin perturbation experiments are complicated by massive interface
movement (fig. 4.7e, fig. 4.9c initially as well as during the experiments, this might represent the 2 time-
scales shown in Chapter 3, Fig. 3.5a). Branched actin networks can be stiff enough to sustain forces
produced during leading edge extension of motile cells or in vitro actin-based propulsion of beads (Marcy
et al. 2004, Chaudhuri et al. 2007). But they can rupture or tear when stretched beyond a certain limit
(Paluch et al. 2006). Particularly for initial offsets of >1μm (cyto D, fig. 4.9 b and c, latrunculin A
observed but not shown), interface movement can exceed 1 μm. This initial movement clearly implies
that the trap does produce considerably more force on the cell-cell interface than in water-injected
embryos. However interface amplitude was not dependent on interface movement (data not shown). The
existence of 2 responses to cytochalasin (cyto D small and large, fig. 4.9 b and c) might shed some light
on it. Particularly for large initial offsets in fig. 4.9 c, the interface moves massively in both scenarios but
75
seen in germband extension. While cadherin intensity in germband extension correlates inversely with
interface stiffness, dissecting the role of cadherin from that of myosin is not straightforward. To
investigate any apparent causal relation between cadherin organization and interface tension would
require further experiments that perturb the cadherin levels directly (for example alpha-catenin RNAi).
Given that time dependence of interface amplitude does not change from stage 5 to stage 7
(Chapter 3 and Fig. 4.9 d), this implies that viscous damping remains unchanged. Therefore the
apparent increase in phase lag at stage 7 (Fig. 4.8a) arises from another mechanism, possibly
from the tissue movement induced by germband elongation. This is further indicated from the
observation that injection with ROCK, cytochalasin D or latrunculin A blocked germband extension
(data not shown) in my experiments and exhibit similar phase lag between trap and interface to stage 5
(fig. 4.8a). The massive increase in interface movement, amplitude and failure to recover intial
position, observed in cyto D large and small (Fig. 4.7e, Fig. 4.9 b and c) imply a creep-like behavior.
Most of the movement is localized at the point of trap-interface contact (Fi.g 4.9 b and c snapshots)
compared to wildtype where movement spreads across the interface length (Fig. 4.9a snapshot). This
might be due to clusters of higher actin density produced by cytochalasin D along with cortex
disruption (Fig. 4.7a). Large movements are also observed in latrunculin A despite small average
amplitude, though not to the extent seen in cytochalasin D (Fig. 4.7e, Fig. 4.7d). Since the interface is far
from its initial position in these measurements, the behavior may not be quasistatic and direct
interpretation of tension from deflection is not possible. In this scenario, the pull and release experiment
(Fig. 3.5a) would be more suitable, both to understand interface tension and creep (especially over
longer time-scales). We do see some change in interface amplitude for successive cycles at the same
time period in cyto D large (Fig. 4.9c). Whether this indicates sinusoidal cyclic loading or is a
consequence of the constantly moving interface (Fig. 4.9c) is unclear.
76
4.3.3 Creep and sinusoidal loading
the response is faster in cyto D large. After this point the interface shows either large or small deflection
amplitudes. It might be that in cyto D large, the cortex is stretched beyond its limit, whereas in cyto D
small it continues to resist deformation.
4.4 Summary
Here I showed that optical tweezers can be robustly used to measure variations in tension at cell contacts
arising from tissue movements in vivo, and that this method can reveal planar-polarized as well as
apicobasal anisotropies of tension in a developing organism. The role of myosin dynamics in establishing
this polarity is established further from my results. I also observed an apico-basal polarization of tension
with changing cadherin levels in the absence of myosin dynamics, although the causality of this
correlation was not established. The experiments here, furthermore, confirmed that the actin cortex is
responsible for interface tension, especially in the absence of cadherin junctions and myosin
activity. Actin disruption can induce creep-like behavior in interface mechanics. Further work in
this direction, particularly with push and pull experiments will help in establishing a dynamic
viscoelastic model.
77
78
CHAPTER 5
CONCLUSIONS AND PERSPECTIVES
79
In this thesis, I have studied the mechanical properties of cell-cell contacts during tissue morphogenesis
and how local deformation propagates within a tissue. This work provides absolute values of tensions at
cell interfaces, while previous work estimated relative values based on assumptions of the viscous
properties at cell junctions. Surprisingly, the small forces produced by optical tweezers are sufficient to
produce significant deflection of cell interfaces and the tension at cell interfaces was estimated to be in
the range of tens to hundreds of pN during early stages of tissue morphogenesis. This suggests that the
forces that remodel cell-cell contacts during tissue morphogenesis and drive the shrinkage or extension of
cell contacts can be powered by a small number of molecular motors. Fluctuations in cell shape, which
are observed during these events, might thus result from stochastic fluctuations in motor numbers.
My work allowed designing a predictive mechanical model of cell contacts. Mechanical modeling of
epithelia is crucial for understanding epithelial morphogenesis events; suggested as early as 1981 by
Odell’s pioneering work (Odell et al. 1981). Since then, a variety of mechanical descriptions have been
proposed; however, testing of the underlying hypotheses has been limited due to the lack of in
vivo experimental tools. Notably, so-called vertex models (Fletcher et al. 2014), usually based on
energy minimization, do not incorporate energy dissipation and thus cannot predict the tissue
dynamics. Here a model is proposed, which does incorporate a viscoelastic constitutive behavior.
Therefore, this model incorporates the vertex models and more recent continuum mechanics
approaches (Bonnet et al. 2012, Hutson et al. 2009). In addition, it captures the non-trivial two-
timescale relaxation dynamics evidenced by pull-release experiments. The possibility of absolute
tension measurements at cell contacts might be beneficial to force inference methods (Chiou et al. 2012,
Cranston et al. 2010, Ishihara et al. 2013), which provide relative tensions based on the geometry
of the contact network. It might indeed allow experimental validation of the inference, but also
be used to calibrate the inferred tensions. The analysis was intentionally restricted to time scales and
speeds faster than the changes in contractility in order to deal with steady shape patterns. The
approach established here could allow the exploration of additional time scales and to probe long-term
plastic deformation of cell contacts.
One important aspect that cannot be overlooked is that our model is passive (mainly incorporating stage 5
end, a tissue with minimal acto-myosin activity or morphogenetic movements) and does not incorporate
active tissue properties such as the role of acto-myosin. Extending mechanical understanding to, for
example, germband elongation, was the motivation behind experiments in chapter 4. Particularly the
creep-like, actin dependent behavior that I observed needs to be studied carefully using push-pull
experiments. Simultaneous imaging of acto-myosin as well as cadherin dynamics in such experiments
could provide the crucial data missing for developing an active model.
80
Lastly, the trapping behavior needs to be better understood, especially the biological origin of refractive
index mismatch at the cell-interface. Understanding this will help better to extend this approach as a
general system-independent tool for understanding the mechanics of morphogenesis. Such approaches are
being developed in the lab currently. Particularly we have replicated the trapping behavior in different
optical and imaging setups (inverted microscope with spinning disk, optical tweezers with SLM) as
well as in another model system (C. elegans embryo). Additionally, having characterized the behavior
for a single trap, this analysis can be extended to novel configurations, for example multiple traps as
well to a Spinning disk setup equipped with a laser ablation system will provide better mechanical
understanding of morphogenetic events. The availability of FRET-based force sensors in Drosophila,
as well as their previous usage in literature (Cai et al. 2014) in conjunction with the approaches being
developed, provide an exciting perspective on the advancements to come in this emerging field that I
have advanced and contributed to.
81
ACKNOWLEDGEMENTS
I
This thesis has been an important part of my journey in understanding life as well as my
directions and goals over the last few years. I arrived in France 3 years ago with no knowledge
of French or Europe, as well as a crippling mood disorder (which was later re-diagnosed as
Aspergers). I owe much to Pierre-François for helping me through all the initial paperwork as well
as being patient with my chaotic and sloppy method of working. The various meetings with him
have been inspiring and motivational even if I had my own way of doing things. I'm grateful to
Olivier Blanc, Claire and Raphael for all the help with the work and data analysis, particularly the
programming which I never developed an interest in, until recently. Sebastien, for all the help
with the fly strains as well as biochemical protocols. All the members of the lab for providing a
helpful and safe environment to work in as well as the occasional bursts of non-scientific
conversations which have inspired me much in my artistic pursuit. IBDML has been a very good
environment to work, nestled in the forest of Luminy and its captivating environment. All the
members of the Lecuit team have been very helpful during the last few years, discussions with
Manos, Loic and Claudio were in particular, very insightful in the understanding and progress of my
work. Willi, Diego, Pierluigi, Sabrina have been good friends and support systems in addition to
all the scientific discussions. Outside of the department, the LABEX initiative and the various
interactions that it provided , has been very positive. Discussions with Pierre-Henri Puesch on Linux,
mindfulness and art have been life-changing and certainly play an important role in my future
pursuits. This thesis is dedicated to all my family of soul in the South of France. All my friends
from CCL for the helping hand during my initial social awkwardness. Hugo, Jess and Nico
for the artistic collaborations and much needed support during thesis writing. Cortical Systematics
Crew for helping me heal my paranoia and mistrust of people through the various events they
organized and I had the good fortune to be part of. And to all the reflections that I crossed my
paths with either in my journeys or through discovery of their work.
REFERENCES
Chapter 1
Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson JD. (2002). Molecular Biology of
the Cell, Garland Science Publ., NewYork.
Ashkin A. (1997) Optical trapping and manipulation of neutral particles using lasers. Proc. Natl.
Acad. Sci. USA 94:4853–4860.
Ashkin A., Dziedzic JM. (1989). Internal cell manipulation using laser traps. Proc. Natl. Acad. Sci.
USA 86:7914–7918.
Baum B., Georgiou M. (2011) Dynamics of adherens junctions in epithelial establishment,
maintenance and remodeling. J. Cell Biol. 192(6):907–917.
Bertet C., Sulak L., Lecuit T. (2004) Myosin-dependent junction remodelling controls planar cell
intercalation and axis elongation. Nature 429:667–671.
Bonnet I., Marcq P., Bosveld F., Fetler L., Bellauche Y., Graner F. (2012). Mechanical state,
material properties and continuous description of an epithelial tissue. arXiv:1205.2454.
Bray D. (2000) Cell Movements–From Molecules to Motility, Garland Science Publ., NewYork.
Burke-Smith A. Cellular organization of tissues. MCD Tissues.
Burridge K., Chrzanowska-Wodnicka M. (1996) Focal adhesions, contractility, and signaling,
Annu. Rev. Cell Dev. Biol. 12 :463.
Cai D., Chen SC., Prasad M., He L., Wang X., Choesmel-Cadamuro V., Sawyer JK., Danuser G.,
Montell DJ. (2014) Mechanical feedback through E-cadherin promotes direction sensing
during collective cell migration. Cell 157:1146–1159.
Cavey M., Rauzi M., Lenne PF., Lecuit T. (2008) A two-tiered mechanism for stabilization
and immobilization of E-cadherin. Nature 453:751-757.
Daniels BR., Masi BC., Wirtz D. (2006) Probing single-cell nicromechanics in vivo: the
microrheology of C. elegans developing embryos Biophys. J. 90:4712-4719.
Davidson L., Dassow MV., Zhou J. (2009) Multi-scale mechanics from molecules to morphogenesis.
Int J Biochem Cell Biol. 41(11): 2147–2162.
Fletcher AG., Ostefield M., Baker RE., Shvartsman SY. (2014) Vertex models of
epithelial morphogenesis. Biophys. J. 106:2291-2304.
II
Ghibaudo M., Saez A., Trichet L., Xayaphoummine A., Browaeys J., Silberzan P., Buguin A., and
Ladoux B. (2008) Traction forces and rigidity sensing regulate cell functions. Soft Matter
4(9):1836–1843.
Gorfinkiel N., Blanchard GB. (2011) Dynamics of actomyosin contractile activity during epithelial
morphogenesis. Current Opinion in Cell Biology 23:531–539.
Grashoff C., Hoffman BD., Brenner MD., Zhou R., Parsons M., Yang MT., McLean MA., Sligar SG.,
Chen CS., Ha T. (2010) Measuring mechanical tension across vinculin reveals regulation of focal
adhesions dynamics. Nature 466:263-266.
Grumbling G, Strelets V. (2006) FlyBase: anatomical data, images and queries. Nucleic Acids Res.
34:D484–8.
Harris AR., Daeden A., Charras GT. (2014) Formation of adherens junctions leads to emergence of a
tissue-level tension in epithelial monolayers. Journal of Cell Science 127:2507-2517.
He B., Doubrovinski K., Polyakov O., Wieschaus E. (2014) Apical constriction drives tissue-
scale hydrodynamic flow to mediate cell elongation. Nature 508:392–396.
Hormeno S., Arias-Gonzalez JR. Exploring mechanchemical processes in the cell with optical
tweezers. (2006) Biol. Cell. 98:679–695.
Hutson MS., Tokutake Y., Chang MS., Bloor JW., Venakides S., Kiehardt DP., Edwards GS.
(2003) Forces for morphogenesis investigated with laser microsurgery and quantitative modeling.
Science 300:145–149.
Ishihara S., Sugimara K., Cox SJ., Bonnet I., Bellauche Y., Graner F. (2013). Comparative study of
non-invasive force and stress inference methods in tissue. Eur. Phys. J. E Soft Matter 36:9859.
Kumar A., Shivshankar GV. (2012) Mechanical force alters morphogenetic movements and
segmental gene expression patterns during Drosophila embryogenesis. PLoS. one 7(3):e33089.
Kuo JC. (2013) Mechanotransduction at focal adhesions: integrating cytoskeletal mechanics in
migrating cells. J. Cell. Mol. Med. 17:704-712.
Ladoux B., Anon E., Lambert M., Rabodzey A., Hersen P., Buguin A., Silberzan P., Mege RM. (2010)
Strength dependence of cadherin-mediated adhesions. Biophys. J. 98:534-542.
Lecuit T., Lenne PF., Munro E. (2011) Force generation, transmission, and integration during cell and
tissue morphogenesis. Annu Rev Cell Dev Biol 27:157–184.
Lecuit T., Wieschaus E. (2000) Polarized insertion of new membrane from a cytoplasmic reservoir
during cleavage of the Drosophila embryo. J. Cell. Biol. 150(4):849-860.
Leidel, C., Longoria, R.A., Gutierrez, F.M., Shubeita, G.T. (2012) Measuring molecular motor forces
in vivo: Implications for tug-of-war models of bidirectional transport. Biophys. J. 103(3):492-500.
III
Lim CT., Zhoua EH., Quek ST. (2006) Mechanical models for living cells—a review. Journal of
Biomechanics 39:195-216.
Munoz JJ., Barrett K., Miodownik M. (2007) A deformation gradient decomposition method for
the analysis of the mechanics of morphogenesis. J Biomech 40: 1372–1380.
Odell GM., Oster G., Alberch P., Burnside B. (1981) The mechanical basis of
morphogenesis. Developmental Biology 85:446-462.
Panorchan P., Lee JSH., Daniels BR., Kole TP., Tseng Y., Wirtz D. (2007) Probing cellular
mechanical responses to stimuli using ballistic intracellular nanorheology. Methods in cell biology
8:115-140.
Perry JK., Lins RJ, Lobie PE and Mitchell MD. (2010) Regulation of invasive growth: similar
epigenetic mechanisms underpin tumour progression and implantation in human pregnancy. Clinical
Science (2010) 118:451–457.
Pullarkat PA., Fernandez PA., Albrecht O. (2007) Rheological properties of the eukaryotic cell
cytoskeleton. Physics Reports 449:29– 53.
Rauzi M., Lenne PF., Lecuit T. (2010) Planar polarized actomyosin contractile flows control epithelial
junction remodelling. Nature 468:1110–1114.
Rauzi M., Lenne PF. (2011) Cortical forces in cell shape changes and tissue morphogenesis. Current
Topics in Developmental Biology 95:93-144.
Rauzi M., Verant P., Lecuit T., Lenne PF. (2008) Nature and anisotropy of cortical forces orienting
Drosophila tissue morphogenesis. Nat. Cell. Biol. 10:1401–1410.
Rodriguez M., McGarry PJ., Sniadecki NJ. (2013) Review on cell mechanics: Experimental and
modeling approaches Applied Mechanics Reviews 65 (060801):1-41.
Solon J., Kaya-Copur A., Colombelli J., Brunner D. (2009) Pulsed forces timed by a ratchet-like
mechanism drive directed tissue movement during dorsal closure. Cell 137:1331–1342.
Truong Quang BA., Mani M., Markova O., Lecuit T., Lenne PF. (2013) Principles of e-cadherin
supramolecular organization in vivo. Current biology 23:2197–2207.
Wu PH., Hale CM., Chen WC., Lee JS., Tseng Y., Wirtz D. (2012) High-throughput ballistic injection
nanorheology to measure cell mechanics. Nat. Protoc. 5;7(1):155-70.
Wyczalkowski MA., Chen Z., Filas BA., Varner VD., Taber LA. (2012) Computational models for
mechanics of morphogenesis. J. Birth Defects Res. C Embryo Today 96(2):132–152.
Yonemura S. (2011) Cadherin-actin interactions at adherens junctions. Current opinions in cell
biology 23:515-522.
Zhu C., Bao G., Wang N. (2000) Cell mechanics: mechanical response, cell adhesion, and molecular
deformation. Annu. Rev. Biomed. Eng. 2:189–226.
IV
Chapter 2
Chardès C., Ménélec P., Bertrand V., Lenne PF. (2014) Setting-up a simple light sheet microscope
for in toto imaging of C. elegans development. J. Vis. Exp. 87:e 51342.
Ebbeson CL., Bruus H. (2012) Analysis of laser induced heating in optical neuronal guidance. J.
Neuroscience Methods 209:168– 177.
Huisken J., Swoger J., Del Bene F., Wittbrodt J., Stelzer EH. (2004) Optical sectioning deep inside
live embryos by selective plane illumination microscopy. Science 305:1007–1009.
Chapter 3
Alenghat FJ., Fabry B., Tsai KY., Goldmann WH., Ingber DE. (2000) Analysis of cell mechanics
in single vinculin-deficient cells using a magnetic tweezer. Biochemical and biophysical
research communications 277:93–99.
Ashkin A. (1997) Optical trapping and manipulation of neutral particles using lasers. Proc.
Natl. Acad. Sci. USA 94:4853–4860.
Ashkin A., Dziedzic JM. (1989) Internal cell manipulation using laser traps. Proc. Natl. Acad. Sci.
USA 86:7914–7918.
Bar-Ziv R., Moses E., Nelson P. (1998) Dynamic excitations in membranes induced by
optical tweezers. Biophys. J. 75:294-320.
Bon P., Maucort G., Wattellier B., Monneret S. (2009) Quadrilateral shearing interferometry
for quantitative phase microscopy of living cells. Optics Express 17:13080-13094.
Welte MA., Gross SP., Postner M., Block SM., Wieschaus EF. (1998) Developmental regulation
of vesicle transport in Drosophila embryos: Forces and kinetics. Cell 92:547-557.
He B., Doubrovinski K., Polyakov O., Wieschaus E. (2014) Apical constriction drives tissue-
scale hydrodynamic flow to mediate cell elongation. Nature 508:392–396.
Maruthamuthu V., Sabass B., Schwarz US., Gardel ML. (2011) Cell-ECM traction force
modulates endogenous tension at cell-cell contacts. Proc. Natl. Acad. Sci. USA 108:4708–4713.
Mayer M., Depken M., Bois JS., Julicher F., Grill SW. (2010) Anisotropies in cortical tension reveal
the physical basis of polarizing cortical flows. Nature 467:617-621.
V
Chapter 4
Bendix PM., Koenderink GH., Cuvelier D., Dogic Z., Koeleman BN. (2008). A quantitative
analysis of contractility in active cytoskeletal protein networks. Biophys. J. 94:3126–36.
Bertet C., Sulak L., Lecuit T. (2004) Myosin-dependent junction remodelling controls planar
cell intercalation and axis elongation. Nature 429:667–671.
Blankenship JT., Backovic ST., Sanny JS., Weitz O., Zallen JA. (2006) Multicellular rosette
formation links planar cell polarity to tissue morphogenesis. Dev Cell. 4:459-70.
Cavey M., Rauzi M., Lenne PF., Lecuit T. (2008) A two-tiered mechanism for stabilization
and immobilization of E-cadherin. Nature 453:751-757.
Chaudhuri O., Parekh SH., Fletcher DA. (2007) Reversible stress softening of actin networks.
Nature 445:295–98.
Chen H., Cohen DM., Choudhury DM., Kioka N., Craig SW. (2005) Spatial distribution and
functional significance of activated vinculin in living cells. J. Cell Biol. 169:459–70.
Harris TJC., Peifer M. (2004) Adherens junction-dependent and -independent steps in
the establishment of epithelial cell polarity in Drosophila. J. Cell. Biol. 167:135–147.
Kölsch V., Seher T., Fernandez-Ballester GJ., Serrano L., Leptin M. (2007) Control of
Drosophila gastrulation by apical localization of adherens junctions and RhoGEF2. Science
315:384–386.
Lecuit T., Lenne PF., Munro E. (2011) Force generation, transmission, and integration during cell
and tissue morphogenesis. Annu. Rev. Cell. Dev. Biol. 27:157–184.
Marcy Y., Prost J., Carlier MF., Sykes C. (2004) Forces generated during actin-based propulsion: a
direct measurement by micromanipulation. Proc. Natl. Acad. Sci. USA 101:5992–97.
Paluch E., Van der Gucht J., Sykes C. (2006) Cracking up: symmetry breaking in cellular systems. J.
Cell Biol. 175:687–92.
Rauzi M., Verant P., Lecuit T., Lenne PF (2008) Nature and anisotropy of cortical forces
orienting Drosophila tissue morphogenesis. Nat. Cell. Biol. 10:1401–1410.
Truong Quang BA., Mani M., Markova O., Lecuit T., Lenne PF. (2013) Principles of e-
cadherin supramolecular organization in vivo. Current biology 23:2197–2207.
VI
Chapter 5
Bonnet I., Marcq P., Bosveld F., Fetler L., Bellauche Y., Graner F. (2012) Mechanical state,
material properties and continuous description of an epithelial tissue. arXiv:1205.2454.
Cai D., Chen SC., Prasad M., He L., Wang X., Choesmel-Cadamuro V., Sawyer JK., Danuser G.,
Montell DJ. (2014) Mechanical feedback through E-cadherin promotes direction sensing
during collective cell migration. Cell 157:1146–1159.
Chiou KK., Hufnagel L., Shraiman BI. (2012) Mechanical stress inference for two dimensional
cell arrays. PLoS. computational biology 8:e1002512.
Cranston PG., Veldhuis JH., Narasimhan S., Brodland GW. (2010) Cinemechanometry (CMM):
A method to determine the forces that drive morphogenetic movements from time-lapse images.
Ann. Biomed. Eng. 38:2937–2947.
Fletcher AG., Ostefield M., Baker RE., Shvartsman SY. (2014) Vertex models of
epithelial morphogenesis. Biophys. J. 106:2291-2304.
Hutson MS., Tokutake Y, Chang MS, Bloor JW, Venakides S, Kiehart DP, Edwards GS.
(2003) Forces for morphogenesis investigated with laser microsurgery and quantitative
modeling. Science 300:145–149.
Ishihara S., Sugimara K., Cox SJ., Bonnet I., Bellauche Y., Graner F. (2013) Comparative study of
non-invasive force and stress inference methods in tissue. Eur. Phys. J. E Soft Matter 36:9859.
Odell GM., Oster G., Alberch P., Burnside B. (1981) The mechanical basis of
morphogenesis. Developmental Biology 85:446-462.
VII
1
Direct laser manipulation reveals the mechanics of cell contacts in vivo
Kapil Bambardekar1,*
, Raphaël Clément1,*
, Olivier Blanc1, Claire Chardès
1 and Pierre-François Lenne
1
1. Aix Marseille Université, CNRS, IBDM UMR7288, 13009, Marseille, France.
Corresponding author: [email protected]
* K.B. and R.C. contributed equally to this work
Abstract
Cell-generated forces produce a variety of tissue movements and tissue shape changes. The
cytoskeletal elements that underlie these dynamics act at cell-cell and cell-extracellular matrix contacts
to apply local forces on adhesive structures. In epithelia, force imbalance at cell contacts induces cell
shape changes, such as apical constriction or polarized junction remodeling, driving tissue
morphogenesis. The dynamics of these processes are well characterized; however, the mechanical
basis of cell shape changes is largely unknown due to a lack of mechanical measurements in vivo. We
have developed an approach combining optical tweezers with light sheet microscopy to probe the
mechanical properties of epithelial cell junctions in the early Drosophila embryo. We show that
optical trapping can efficiently deform cell-cell interfaces and measure tension at cell junctions, which
is on the order of picoNewtons (pN). We demonstrate that tension at cell junctions equilibrates
over few seconds; a short time scale compared to the contractile events that drive morphogenetic
movements. We further show that tension increases along cell interfaces during early tissue
morphogenesis and becomes anisotropic as cells intercalate during germ band extension. By
performing pull-and-release experiments, we identify time-dependent properties of junctional
mechanics consistent with a simple visco-elastic model. Integrating this constitutive law into a tissue-
scale model, we predict quantitatively how local deformations propagate throughout the tissue.
Significance statement
The shaping of tissues and organs relies on the ability of cells to adhere together and to deform in a
coordinated manner. It is therefore key to understand how cell-generated forces produce cell shape
changes, and how such forces transmit through a group of adhesive cells in vivo. In this context, we
have developed an approach using laser manipulation to impose local forces on cell contacts in the
early Drosophila embryo. Quantification of local and global shape changes using our approach can
provide both direct measurements of the forces acting at cell contacts, and delineate the time-
dependent visco-elastic properties of the tissue. The latter provides an explicit relationship, the so-
called constitutive law, between forces and deformations.
2
\body
Introduction
During the development of an organism, cells change their shape and remodel their contacts to give
rise to a variety of tissue shapes. Analysis of tissue kinematics has revealed that epithelial tissue
morphogenesis is partly controlled by actomyosin contractility. The spatio-temporal deployment and
coordination of actomyosin contractility produces shrinkage and extension of cell surfaces and
interfaces, which can drive tissue invagination, tissue folding, or tissue extension (1). Understanding
the mechanical nature of these processes requires force measurements in vivo; however, measurements
in developing epithelia are limited and most methods have been indirect. They rely either on force
inference from image analysis (2-4) or on laser dissection experiments at cell (5, 6) or tissue scales
(7, 8), which provide the relative magnitude and direction of stresses from cell or tissue shape
changes. In contrast, mechanical approaches have been developed in recent years to impose or
measure stresses of cells in contact, including cell monoloayer micromanipulation (9), pipette
microaspiration on cell doublets (10), and traction force microscopy on migrating epithelia (11) and
single cell doublets (12). Recently, an elegant method using deformable cell-sized oil microdroplets
has provided absolute values of stresses at the cell level in cell cultures and embryonic mesenchymes
(13), but not yet in live epithelia. In this context, we sought for a direct, in vivo method for tension
measurements and mechanical characterization at cell contacts, and developed an experimental
approach combining optical tweezers with light sheet microscopy.
To probe epithelial mechanics in a live organism, we chose the early epithelium of the Drosophila
embryo as a model system. It consists of a simple sheet of cells that spread over the yolk and are in
contact with each other through E-cadherin-based adhesion. During early embryogenesis at the
blastula stage just after the end of cellularization, epithelial cells have very similar hexagonal shapes,
suggesting that cell junctions have similar mechanical properties and the internal pressure of these
cells is homogeneous. At the later gastrula stage, cells undergo shape changes at distinct regions in the
embryo. On the ventral side of the embryo, apical cell constriction of a few rows of cells drives tissue
invagination (14), while on the ventrolateral side of the embryo, cell intercalation, a process whereby
cells exchange neighbors by polarized remodeling of their junctions, drives tissue extension. The latter
morphogenetic movement is driven by an anisotropic distribution of Myosin-II (Myo-II), which is
more concentrated along junctions aligned with the dorsal-ventral (D/V) axis (15). Laser dissection of
cortical actomyosin networks at cell junctions in the ventrolateral tissue has shown that such an
anisotropic distribution of Myo-II causes an anisotropic cortical tension (6). However, the absolute
values of tensile forces have not yet been measured, and more generally, the mechanics of cell-cell
interfaces in vivo is largely unknown. Here we addressed this issue by analyzing local mechanical
measurements at cell junctions during tissue morphogenesis and determining the contribution of Myo-
II to tension in this context. We determined the time-dependent response of cell-cell interfaces to
3
forced deflection and delineated a visco-elastic model of junctions. Finally, this led us to explore the
propagation of local forces within the epithelial tissue.
Results and Discussion
To probe the mechanics of cell-cell interfaces, we devised a setup combining optical tweezers and
light sheet microscopy (Fig. 1A and Fig. S1). This combination allows imaging of a whole tissue at a
high acquisition rate, while manipulating objects in vivo. Light sheet microscopy is also advantageous
to confocal microscopy because it reduces photobleaching (15). Our light sheet setup was designed
from an upright microscope: a light sheet sections the sample horizontally and the fluorescence light is
collected by a high numerical aperture objective lens pointing downwards (16) (Fig. 1A and Fig. S1).
The laser trap is produced by a near-infrared laser light focused by the collection objective lens into
the sample and is moved by galvanometric mirrors in the plane of the epithelium (Fig. 1A and Fig.
S1). Optical tweezers experiments usually require the use of a glass or a polystyrene bead to apply a
force onto an attached molecule or a cellular structure. We found that the cell-cell interfaces can be
manipulated directly, without the need of an external probe (Fig. 1B showing 3 snapshots of deflected
membrane interfaces and trap position; Fig. 1C; Movie 1). This is likely due to a positive refraction
index difference between the interface and the interior of the cells, as revealed by quantitative phase
imaging of the epithelial cells (Fig. S2). Yet, the value of this mismatch is difficult to determine
because of the geometry of the tissue.
Using the direct application of the laser, we imposed a sinusoidal movement to the trap perpendicular
to a cell interface and centered on it, and imaged the resulting deflection in the adherens junction plane
(Fig. 1C). The interface deflection followed the trap movement but with lower amplitude, suggesting
that the interface resists the mechanical load imposed by the laser trap (Fig. 1C).
To explore the regime of deformation that the laser trap imposes to the interfaces, we varied the
amplitude of the trap sinusoidal movement while keeping the period of oscillation constant. The
amplitude of the interface deflection increases with the trap amplitude, yet it deviates from a linear
relationship for trap amplitude larger than 1 µm (Fig. 1D). Then, we varied the laser power while
keeping the trap amplitude and the period of oscillation constant. We found that the interface
deflection amplitude also increases linearly with the laser power up to 300 mW (Fig. 1D, inset).
Together, these results confirm that in the case of small deformations (for comparison the average
length of an interface is 4.5 µm), the trap acts as a linear spring, whose stiffness �� is linearly
proportional to the laser power. Therefore, all the experiments are carried out within this range of
deformation (<1 µm trap amplitude) and with power 200 mW, unless otherwise stated.
To estimate the trap stiffness on the cell interfaces, and thus the forces directly applied by the
laser, we implemented a two-step procedure using beads. First, we determined the trap stiffness on
beads: single 0.46 µm diameter beads injected in the cytosol were trapped and moved in a stepwise
fashion between two trap positions separated by 0.5 µm (Fig. S3A). The resulting relaxation of the
4
bead towards the new trap position was exponential. The characteristic time is set by the ratio of the
drag coefficient, � �, over the trap stiffness on the bead, with the viscosity and R the bead radius.
An effective value of the viscosity was estimated independently by analyzing bead motion in the
cytosol in the absence of trap to measure its mean square displacement (Fig. S4) (17). Relating the
diffusion constant to viscosity by the Stokes-Einstein equation (� Θ = � �), where � is the
Boltzmann constant and Θ the temperature, we found that the effective viscosity of the cytosol is
3.6 ± 0.1 Pa.s (mean ± se, 1350 beads). Using this value, we could thus estimate the trap stiffness on
beads to be 120 ± 50 pN.µm-1
at 200 mW laser excitation (mean ± sd, 20 measurements). Second, we
compared the deformation produced by direct application of the focused laser on the interface with
that induced by beads pushed against the cell-cell interfaces (Fig. S3B-D and Movie 2). The former
was only 2- to 3-fold larger than the latter (2.5 ± 0.4, mean ± sd, 5 measurements), indicating that the
trap stiffness on the interfaces was 2- to 3-fold smaller than that on beads. Thus, the trap stiffness on
interfaces was estimated to be �� = 50 ± 30 pN.µm-1 at 200 mW laser excitation and in the regime of
small deformations.
The resistance to deformation can arise not only from the mechanical properties of the
interface and its apposed cortical elements, including the actomyosin cytoskeleton, but also from the
viscous cytosol. To determine whether the resistance to deformation is time-dependent, we varied the
period of oscillation while keeping the trap amplitude constant (amplitude At = 0.5 µm, Fig. 1E). For
periods larger than or equal to 1 s, which correspond to mean speeds smaller than 2 µm.s-1
, the
amplitude was constant. The constancy of the deflection amplitude at speeds below 2 µm.s-1
indicates
that the force applied by the trap and that produced by the interface are in quasi-static equilibrium. At
low speeds of deformation (speed < 2 µm.s-1
), we can thus assume that the shape of the interface
mainly results from the balance of forces between the trapping force �� and the tension of the interface�: �� = � cos , where is the angle that the interface makes with respect to the trapping force (Fig.
1B). As the vertices of the cell-cell contact did not move significantly during the deformation (Fig.
S5), we could neglect the contribution of other cells and use this simple local equilibrium formula.
For small deformations (that is for maximal deflections much smaller than the initial junction length
l0), cos ≈ �0 and �� ≈ �� �� − � , where � is the position of the interface and �� the position of the
trap. The tension of the interface thus approximates as: � ≈ � 04 ��� − . We found that within
experimental error the ratio ��� remains constant during periodic oscillations (Fig. 1F), indicating that
the pre-existing tension � of the interface is not significantly modified during small deformations.
Importantly, this also implies that tension measurements, while relying on geometrical and physical
approximations, do not require a mechanical model of cell contacts. Thus, tension values can be
obtained as a simple linear function of the ratio between the interface deflection and the trap position.
Using our estimate of �� = 50 pN.µm-1
, and that the ratio ��� = 1.88 ± 0.4 (mean ± sd, 16
5
measurements), we found that tension � at cell-cell interfaces is on the order of 44 ± 22 pN at the end
of cellularization. The tensions reported here are in the same range as cortical tensions measured on
single cells (18), but 2-3 orders of magnitude below cell-cell forces in cell aggregates on adhesive
substrates in vitro (12).
We observed that for periods smaller than 1 s (speed > 2 µm.s-1
), the deflection was reduced
(Fig. 1E). This is the characteristic signature of viscous damping, presumably related to the viscous
drag in the cytosol. To test whether damping was indeed due to the viscosity of the cytosol, we
performed relaxation experiments after instantaneous release of the trap (Fig. S6A). At the onset of
relaxation, tension is balanced only by viscous damping: � cos = �� , where � is the initial
relaxation velocity and �is the damping coefficient. Measuring � (Fig. S6A) and using the mean
tension value of 44 pN, this provided an indirect measurement of �, which is in the order of 2 ±
1×10-5
m.Pa.s. To determine viscosity, � should be rescaled by both a typical length scale L of
deformation and a geometric coefficient g: = ���. In the plane of junctions, the deformation extends
to the whole contact line (4-5 µm). We found a similar value for the deformation along the apico-basal
direction (Fig. S6B). The deformation is thus likely to be akin to a two-dimensional Gaussian, with a
typical width of 4-5 µm. Therefore we took � = . µm and = , which corresponds to disk
approximation. The viscosity associated with the observed damping is thus on the order of 1 Pa.s,
which is consistent with our previous measurements of cytosol viscosity using beads. This is also
consistent with microrheological measurements made in the cytosol of C. elegans embryo (19).
During tissue morphogenesis, cells undergo cell shape changes driven by Myo-II contractile events,
which induce interface deformation at various speeds up to about 0.1 µm.s (20, 21). The condition of
low speeds (speed < 2 µm.s-1
) is thus always fulfilled during cell shape changes driven by Myo-II
contractility during tissue morphogenesis of Drosophila, indicating that short-time viscous damping
should play no role in the process.
Using direct optical manipulation, we then probed how cell-cell tensions change during tissue
morphogenesis of the early germband of the Drosophila. Before gastrulation movements, at the end of
stage 5, cells form a regular lattice with isotropic shapes (Fig. 2A, top panels) (6, 22, 23). Later, at the
end of stage 6 and the onset of stage 7, the total concentration of Myo-II increases at adherens
junctions, and its distribution becomes anisotropic, with higher levels along interfaces parallel to the
D/V axis (Fig. 2A, middle and bottom panels) (6, 22, 23). This anisotropic distribution of Myo-II has
been shown to drive polarized junction shrinkage and cell intercalation (6, 22, 23). We found that the
deflection amplitude of cell-cell interfaces caused by the optical trap is reduced by two-fold from the
end of stage 5 to the onset of stage 7, indicating a tension increase (Fig. 2B). Moreover, at stage 7,
cell-cell interfaces with a direction close to the D/V axis are about 2.5 times more tense than those
along the horizontal axis (Fig. 2C and Fig. S7 showing tension normalized to junction size). The
6
anisotropy of tension is consistent with a previous report, which estimated the relative values of
tension from laser nanodissection (6).
Inhibition of Myo-II activity by injection of ROCK inhibitor resulted in a significant reduction of the
tension at cell junctions at stage 7 (Fig. 2D) and a loss of tension anisotropy (Fig. 2C). This confirms
that the significant increase in tension measured between stage 5 and stage 7 can be attributed to Myo-
II activity. Given that a single molecular motor of Myo-II produces a few pN of force (24), the range
of forces measured here suggests that the increase in tensions from stage 5 to stage 7 could be powered
by only a few dozens of motors.
The E-cadherin junctions, where we performed the experiments presented above, are restricted to a
thin 1 µm section and are localized about 1-2 µm below the apical surface (25, 26) (Fig. 2D, left
panel). As Myo-II accumulates at the adherens junction plane at the end of stage 6 and onset of stage
7, we wondered if this might translate into different mechanical properties at the adherens junction
plane as compared to more basal positions. Measurements of interface deflection by laser trap at
different positions along the apicobasal axis showed that during tissue morphogenesis, there is a
gradual polarization of the tension along this axis (Fig 2E, lower panel). While at stage 5, tension at
adherens junctions is the same as in a more basal plane, it becomes larger at stage 7. The fact that we
measure the same deformation at the adherens junction plane and 3 µm more basally, at stage 5 end,
when Myosin-II is very apical and not junctional, also suggests that the apical cortex has not a
significant contribution to the restoring force. Altogether these results show that optical tweezers can
be robustly used to measure tension at cell contacts in vivo, and that this method can reveal planar-
polarized as well as apicobasal anisotropies of tension in a developing organism.
This led us to explore in more detail the mechanical response of cell-cell contacts to forced deflection,
at different temporal and spatial scales. First we performed pull-release experiments, which have been
used in vitro on single cells with optical and magnetic tweezers (27), but have never been applied in
vivo. Pull-release experiments consist of switching the laser trap on/off at a few hundred nanometers
distance from the junction and then monitoring the deflection of the cell-cell interface, both towards
(trap on) and away from the trap (trap off) (Fig. 3A). At these short time scales, the mechanical
response of an epithelial tissue should result both from the constitutive mechanics of its actomyosin
cortex and from viscous friction exerted by the cytosol. In other words, modeling deflection dynamics
requires both a cortical constitutive equation, and a force balance equation between the cortical
restoring force, the trapping force, and the viscous friction. The pull-release curves obtained show that
the dynamics is not purely exponential, and exhibits at least two characteristic times, in the range of
1 s and 10 s (Fig. 3A and Fig. S8A). We considered different types of visco-elastic constitutive models
coupled to the force balance equation to fit the experimental data, and found that a so-called standard
linear solid (SLS) model, composed of a Maxwell arm (a spring and a dashpot) in parallel with a
spring, is the best and simplest model to correctly account for the observed behavior (Fig. 3A; fit:
7
solid blue curve, SLS model: inset). Indeed, we can rule out Kelvin-Voigt and Maxwell models, which
predict simple exponential relaxation (Fig. S8B). Notably, the SLS model also fits well the periodic
experiments described above (Fig. 1E, solid blue curve). In particular, the 1 s time scale is consistent
with the drop of deflection amplitude observed for periods of oscillation below 1 s, which can
therefore be attributed to the viscous drag in the cytosol. The existence of another time scale (10 s)
denotes the fact that the cortex itself is not purely elastic but visco-elastic (hence the SLS model).
These two time scales can be derived analytically, and indeed, one is determined by the friction in the
cytosol, while the other is given by the viscous component of the cortex constitutive equation.
Importantly, both these visco-elastic time scales are under a minute; neither the experiments nor the
model deal with the long-term dynamics (minutes to hours), which presumably implies creep and
therefore fluid-like behavior (9, 28).
During tissue morphogenesis, the integration of local forces shapes the tissue (1). A
challenging question is whether local forces produce long-range deformation, and at what speed the
mechanical information propagates. Thus, having established a model for the mechanics of single cell
interfaces, we then asked how single cell deformation propagates throughout the tissue. We imposed
the local deflection of a cell interface using sinusoidal oscillations, and we tracked the deflection of
neighboring interfaces away from this point. For that purpose, we plotted kymographs along lines
perpendicular to cell interfaces (Fig. 3B). The target interface was oscillated using a deflection
amplitude of 1 ± 0.1 m (Fig. 3C, left panel), which is larger than in the experiments presented so far
in order to facilitate the detection of propagation. We observed that the neighboring interfaces within a
distance of 1 to 2 cells also deflected periodically, but with much lower amplitudes and a small phase
shift (Fig. 3C, left panel and 3D, left panel and Fig. S9A-D). This indicates that the deformation
typically decays over a distance of 1-2 cells.
To evaluate the ability of our simple mechanical model to reproduce these data, we first
transposed the single junction model to a network of contacts, as observed in vivo. Each contact is
considered as an SLS element (again with viscoelastic parameters � , � and ) pulling on its vertices
(Fig. 3A). The displacement of each vertex then results from the force balance between tensile forces
exerted by adjacent contacts, and external damping, ��̇, due to cytosol viscosity (Fig. 3E). We were
then able to simulate the mechanical response of the tissue to periodic manipulations in the extracted
experimental geometry, imposing only the sinusoidal displacement at the midpoint of the target
interface (Fig. 3F).
To quantitatively assess the accuracy of the model, we plotted kymographs in the simulated tissues
along the same lines as in the experiments (Fig. 3C, right panel, and Fig. S9). The parameter values
that we estimated from the single interface experiments faithfully reproduce the tissue-scale
observation: one or two neighboring cells deform away from the source point of deflection (Fig. 3C, D
and G, Movie 3). The propagation is a bit more efficient transverse (Fig. S9C) than perpendicular to
deformation (Fig. 3G). Taking higher or lower viscosity for the cytosol into account leads to shorter
8
and longer propagation distance, respectively, which does not correctly reproduce the observed
behavior (Fig. S9E). Note that our model underestimates the speed of propagation (Fig. 3D), which
suggests that constant volume constrains and/or transmission through the apical cortex may contribute
to propagation. The speed of propagation can be experimentally estimated by the phase delay between
the trapped interface’s deflection and that of its neighbors (Fig. 3D, time shift between black and
magenta curves). At a 1-cell distance (~7 m), we measured a time delay of 150 ± 85 ms (95%
confidence interval), which corresponds to a propagation speed with a typical phase velocity of 45
m.s-1
. This speed is much larger than the speed of actomyosin flows, which are observed in different
systems, including Drosophila and C. elegans (0.1 µm.s-1
) (21, 29).
Conclusion
Here we studied the mechanical properties of cell-cell contacts during tissue morphogenesis and how
local deformation propagates within a tissue. We provide absolute values of tensions at cell interfaces,
while previous work estimated relative values based on assumptions of the viscous properties at cell
junctions. Surprisingly, the small forces produced by optical tweezers are sufficient to produce
significant deflection of cell interfaces, and we could estimate that tension at cell interfaces is in the
100 pN range during early stages of tissue morphogenesis. This suggests that the forces that remodel
cell-cell contacts during tissue morphogenesis and drive the shrinkage or extension of cell contacts can
be powered by a small number of molecular motors. Fluctuations in cell shape, which are observed
during these events, might thus result from stochastic fluctuations in motor numbers.
The possibility of absolute tension measurements at cell contacts might be beneficial to force
inference methods (2-4), which provide relative tensions based on the geometry of the contact
network. It might indeed allow experimental validation of the inference, and can also be used to
calibrate the inferred tensions.
Our study has provided a predictive mechanical model of cell contacts. Modeling the
constitutive mechanics of epithelia by quantifying how forces dynamically cause deformations is
crucial for understanding epithelial morphogenesis events, which was suggested as early as 1981 by
Odell’s pioneering work (30). Since then, a variety of mechanical descriptions have been proposed;
however, testing of the underlying hypotheses has been limited due to the lack of in vivo experimental
tools. Notably, so-called vertex models (31), usually based on energy minimization, do not incorporate
energy dissipation and thus cannot predict the tissue dynamics. Here we propose a vertex-based
model, which bridges usual vertex models and continuum mechanics with finite elements approaches
that integrate visco-elastic constitutive behavior (32, 33). In addition, it captures the non-trivial two-
timescale relaxation dynamics evidenced by pull-release experiments. Finally, here we intentionally
restricted our analysis to time scales and speeds faster than the changes in contractility in order to deal
with steady shape patterns. We believe that the approach we have established here is now ready to
explore additional time scales and to probe long-term, irreversible deformation of cell contacts.
9
Materials and methods
Experiments and data analysis
Optical manipulation was done using a custom-built two-colors (488 and 516 nm) light sheet
microscope (16), coupled with a single beam gradient trap (1070 nm wavelength, ytterbium fiber laser,
IPG). A 100x water-immersion lens (1.1 NA, Nikon) was used for imaging as well as for introducing
the optical trap in the imaging plane. Galvanometric mirrors controlled laser trap position deflection to
produce sinusoidal oscillations or step movements. Prior to every in vivo experiment, the relationship
between galvanometer voltages and laser trap position was calibrated using fluorescent beads
(localization precision of 25 nm). During experiments both images and position of the galvanometers
were simultaneously recorded. Kymographs of interface deflection were extracted from movies along
lines perpendicular to the interfaces and were fitted at each time step by a gaussian to determine the
interface position with subpixel resolution (localization precision of 35 nm).
Quantification of E-cadherin::GFP and Squash::GFP was done in a spinning disk microscope (Perkin-
Elmer) using a 100x oil immersion lens (Nikon).
For details on sample preparation, see SI Materials and Methods.
Model
- In the single junction model (Fig. 3A and fit Fig. 1E), the horizontal restoring force is related to the
deflection � of the interface through the SLS constitutive mechanics of the cortex: ̇ + � = � + � �̇ + � � �.k1 and k2 are elastic parameters (N.m
-1), a viscous parameter (m.Pa.s), and the dot denotes a temporal
derivative. The force balance at the interface simply reads: = �� �� − � − ��̇,
where �� and �� are the stiffness and position of the optical trap, and � is the damping coefficient in
the cytosol. Combining these two equations yields two characteristic time scales, one related to
(viscous component of the cortex) and the other to � (damping coefficient of the cytosol).
- In the tissue scale model, each bond has a visco-elastic dynamics, given by the same model: �̇ + � � = � + � �̇ + � + � �,where � is the tension and � is the elongation. At vertex i, tensions of adjacent interfaces (j=adj(i)) are
balanced only by viscous damping. The force balance thus writes:
��i⃗⃗⃗ ̇ = ∑ �ij ⃗⃗ ⃗⃗ =�� , which provides direct access to vertices displacements. The midpoint of the target interface is treated
as a 2-way vertex in the simulations. Its movement is imposed as to mimic the considered experiment.
10
Acknowledgements
We thank Serge Monneret (Institut Fresnel Marseille) for quantitative phase imaging experiments. We
thank Sébastien Sénatore and Edith Laugier for help with bead injection. We thank all members of the
Lenne and Lecuit laboratories for discussions and comments on the manuscript. We also acknowledge
members of the Labex INFORM for discussions (ANR-11-LABX-0054) and the France-BioImaging
infrastructure (ANR-10-INSB-04-01, call "Investissements d'Avenir"). This work was supported by
the Fondation pour la Recherche Médicale (Equipe FRM DEQ20130326509) and an ANR-Blanc grant
(Morfor, ANR-11-BSV5-0008) to P-F.L. K.B. was supported by a PhD fellowship from the Région
PACA & NIKON, O.B. by a PhD MESR fellowship and a FRM individual grant.
11
References
1. Lecuit T, Lenne P-F, Munro E (2011) Force generation, transmission, and integration during
cell and tissue morphogenesis. Annu Rev Cell Dev Biol 27:157–184.
2. Cranston PG, Veldhuis JH, Narasimhan S, Brodland GW (2010) Cinemechanometry (CMM):
A method to determine the forces that drive morphogenetic movements from time-lapse
images. Ann Biomed Eng 38:2937–2947.
3. Chiou KK, Hufnagel L, Shraiman BI (2012) Mechanical stress inference for two dimensional
cell arrays. PLoS computational biology 8:e1002512.
4. Ishihara S et al. (2013) Comparative study of non-invasive force and stress inference methods
in tissue. Eur Phys J E Soft Matter 36:9859.
5. Farhadifar R, Roper JC, Aigouy B, Eaton S, Julicher F (2007) The influence of cell mechanics,
cell-cell interactions, and proliferation on epithelial packing. Curr Biol 17:2095–2104.
6. Rauzi M, Verant P, Lecuit T, Lenne P-F (2008) Nature and anisotropy of cortical forces
orienting Drosophila tissue morphogenesis. Nat Cell Biol 10:1401–1410.
7. Hutson MS et al. (2003) Forces for morphogenesis investigated with laser microsurgery and
quantitative modeling. Science 300:145–149.
8. Bosveld F et al. (2012) Mechanical control of morphogenesis by Fat/Dachsous/Four-jointed
planar cell polarity pathway. Science 336:724–727.
9. Harris AR et al. (2012) Characterizing the mechanics of cultured cell monolayers. Proc Natl
Acad Sci USA 109:16449–16454.
10. Maître J-L et al. (2012) Adhesion Functions in Cell Sorting by Mechanically Coupling the
Cortices of Adhering Cells. Science 338:253–256.
11. Trepat X et al. (2009) Physical forces during collective cell migration. Nature Physics 5:426–430.
12. Maruthamuthu V, Sabass B, Schwarz US, Gardel ML (2011) Cell-ECM traction force
modulates endogenous tension at cell-cell contacts. Proc Natl Acad Sci USA 108:4708–4713.
13. Campàs O et al. (2013) Quantifying cell-generated mechanical forces within living embryonic
tissues. Nature methods 11:183–189.
14. Kölsch V, Seher T, Fernandez-Ballester GJ, Serrano L, Leptin M (2007) Control of Drosophila
gastrulation by apical localization of adherens junctions and RhoGEF2. Science (New York,
NY) 315:384–386.
15. Huisken J, Swoger J, Del Bene F, Wittbrodt J, Stelzer EH (2004) Optical sectioning deep
inside live embryos by selective plane illumination microscopy. Science 305:1007–1009.
16. Chardès C, Ménélec P, Bertrand V, Lenne P-F (2014) Setting-up a simple light sheet microscope for in toto imaging of C. elegans development. Journal of visualized experiments: JoVE 87:e51342.
17. Kusumi A, Sako Y, Yamamoto M (1993) Confined lateral diffusion of membrane receptors as
studied by single particle tracking (nanovid microscopy). Effects of calcium-induced
12
differentiation in cultured epithelial cells. Biophys J 65:2021–2040.
18. Salbreux G, Charras G, Paluch E (2012) Actin cortex mechanics and cellular morphogenesis.
Trends in cell biology 22:536–545.
19. Daniels BR, Masi BC, Wirtz D (2006) Probing single-cell micromechanics in vivo: the
microrheology of C. elegans developing embryos. Bioph J 90:4712–4719.
20. Martin AC, Kaschube M, Wieschaus EF (2009) Pulsed contractions of an actin-myosin
network drive apical constriction. Nature 457:495–499.
21. Rauzi M, Lenne P-F, Lecuit T (2010) Planar polarized actomyosin contractile flows control
epithelial junction remodelling. Nature 468:1110–1114.
22. Bertet C, Sulak L, Lecuit T (2004) Myosin-dependent junction remodelling controls planar cell
intercalation and axis elongation. Nature 429:667–671.
23. Blankenship JT, Backovic ST, Sanny JSP, Weitz O, Zallen JA (2006) Multicellular rosette
formation links planar cell polarity to tissue morphogenesis. Dev Cell 11:459–470.
24. Howard J (2001) Mechanics of motor proteins and the cytoskeleton (Sinauer Associates).
25. Harris TJC, Peifer M (2004) Adherens junction-dependent and -independent steps in the
establishment of epithelial cell polarity in Drosophila. The Journal of Cell Biology 167:135–147.
26. Truong Quang B-A, Mani M, Markova O, Lecuit T, Lenne P-F (2013) Principles of e-cadherin
supramolecular organization in vivo. Current biology: CB 23:2197–2207.
27. Alenghat FJ, Fabry B, Tsai KY, Goldmann WH, Ingber DE (2000) Analysis of cell mechanics
in single vinculin-deficient cells using a magnetic tweezer. Biochemical and biophysical
research communications 277:93–99.
28. He B, Doubrovinski K, Polyakov O, Wieschaus E (2014) Apical constriction drives tissue-
scale hydrodynamic flow to mediate cell elongation. Nature 508:392–396.
29. Mayer M, Depken M, Bois JS, Jülicher F, Grill SW (2010) Anisotropies in cortical tension
reveal the physical basis of polarizing cortical flows. Nature 467:617–621.
30. Odell GM, Oster G, Alberch P, Burnside B (1981) The mechanical basis of morphogenesis. I.
Epithelial folding and invagination. Developmental biology 85:446–462.
31. Fletcher AG, Osterfield M, Baker RE, Shvartsman SY (2014) Vertex models of epithelial
morphogenesis. Biophys J 106:2291–2304.
32. Hutson MS et al. (2009) Combining laser microsurgery and finite element modeling to assess
cell-level epithelial mechanics. Biophys J 97:3075–3085.
33. Bonnet I et al. (2012) Mechanical state, material properties and continuous description of an
epithelial tissue. J R Soc Interface 9:2614–2623.
13
Figure Legends
Figure 1: Characterizing the deflection of cell-cell interfaces imposed by optical tweezers.
(A) Schematic of the setup: the embryo is optically sectioned by a light sheet and imaged while a laser
trap (red) allows manipulation. Top image shows the epithelium labeled by a membrane marker
(GAP43::mcherry) and the laser trap position (yellow arrowhead). (B) Separate images of the interface
in three different positions of deflection (yellow arrowheads, 3 top images). The bottom image is a
merge of the three upper images (position 1 in green, position 2 in red, position 3 in blue). Also a
schematic of deflection with distribution of forces labeled. Scale bar, 5 µm. (C) Representative plot of
deflection versus time showing both trap (red solid line) and interface positions (black solid line). (D)
Amplitude of interface deflection as a function of trap movement amplitude and laser trap power
(inset). 7-13 independent measurements per data point. Error bars represent one standard deviation.
(E) Interface deflection amplitude over the trap oscillation period. (F) Interface position as a function
of trap position during few cycles of laser oscillation (amplitude: 0.5 µm, period: 2 s). Successive
cycles are in different colours (black: first, red: second, green: third). The blue line represents a linear
fit.
Figure 2: Tension at cell contacts before and during germband elongation in wild type and
perturbed embryos.
(A) Left panels: Schematic of tissue elongation and images at three different stages: before elongation
(stage 5 end), at the onset of Myosin-II accumulation at cell junctions (stage 6 end) and during tissue
extension (stage 7). Right panels: images showing the cell interface (purple, GAP43::mcherry), E-
cadherin (green) and Myosin-II (red, MRLC: regulatory light chain of Myosin-II) at the different
stages. (B) Interface deflection amplitude (gray bars) and tension (red bars) at the adherens junction
plane, for different stages (16, 24 and 14 different interfaces measured, respectively). (C) Interface
deflection amplitude at different stages for different junction orientation, along and perpendicular to
the D/V axis. (D) Interface deflection amplitude (gray bars) and tension (red bars) in the adherens
plane in ROCK inhibitor-injected embryos and control embryos (WT: 14 interfaces, water injected; 11
interfaces, ROCK inhibitor: 15 interfaces. Error bars represent one standard deviation. (E) Ratio
between the interface deflection at the adherens junction plane and in a more basal plane (3 µm below
adherens junctions) at different stages. The red line is the median, the box edges are the lower and
upper quartiles, and the whiskers display the total range of measurements. T-test (* not significant,
**p-value <0.01, *** p-value <0.001). Scale bar, 10 µm.
Figure 3: Mechanical model of the interface and tissue response.
(A) Deflection of the interface in a pull-release (trap on - trap off) experiment. The model (blue line)
accurately fits the experimental data (black). The simplest analogous visco-elastic model is a maxwell
14
arm in parallel with a spring (inset). For a trap stiffness kt = 50 pN.m-1
, fit parameters values are:k = 15 pN.m-1
, k = 55 pN.m-1
, ζ = 1.5 × 10-4
m.Pa.s, and Cη = 1.5 × 10-5
m.Pa.s. Note that this
value of Cη is consistent with our previous estimate (2 ± 1 × 10-5
m.Pa.s). The same parameter values
are used in the simulations. (B) Deflection perpendicular to the interfaces is tracked over time along
lines perpendicular to cell interfaces (red), which allows measuring deformation away from the
targeted interface (yellow arrowhead). Scale bar, 10 µm. (C) Kymograph of interface deflections in the
experimental (left) and simulated (right) tissues. Only the interfaces adjacent to the target interface
display significant deflection. (D) Deflection of the target (black) and neighbor interfaces (at 1-cell
and 2-cell distance, magenta and green, respectively) in the experimental (left) and simulated (right)
tissues. (E) In the model, the movement of a vertex results from a balance between tension from
adjacent bonds and viscous friction. (F) Overlay of the undeformed (purple) and deformed (green)
tissue in the experimental (left) and simulated (right) tissue. (G) Spatial decay of interface deflections
over the neighboring cells. Comparison between experiments and simulations.
Position 1 Position 2 Position 3
0 0.2 0.4 0.6 0.8 1 1.20
0.1
0.2
0.3
0.4
0.5
0.6
A C
Time (s)
Inte
rface d
eflection (μm
) Trap
Interface
0 2 4 6 8 10 12
0
1
E
Inte
rface d
eflection a
mplit
ude (μm
)
Period of oscillation (s)
F
Trap deflection amplitude (μm)
Inte
rface
deflection a
mplit
ude (μm
)
00.10.20.30.40.5
0 100 200 300 400
Deflection
(μm
)
Laser Power (mW)
B
D
Trap position (μm)
Inte
rface p
ositio
n (μm
)
Embryo
0 1 2 3 4 50
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.5 1
0
0.5
A
EC D apical
basal
AJ
Stage 7
water inj. ROCK inh.
St. 5 end St. 7
E-cad MRLC
MR
LC
E-cadGAP43 MRLC
Stage 6 end
Stage 7
0
0.05
0.1
0.15
0.2
0.25
0.3
0
50
100
150
200
250
300
Tensio
n(p
N)
Inte
rface D
eflection a
mplit
ude (μm
) ******B
0
50
100
150
200
250
300
350
400
Tensio
n (
pN
)
St. 6 end
St. 5 end St. 7 St. 6 end
PA
D
VP
**
St.7 St.7
water inj.
St.7 ROCK Inh.0
0.05
0.1
0.15
0.2
0.25
0.3
0
50
100
150
200
250
300
Tensio
n(p
N)
Inte
rface D
eflection a
mplit
ude (μm
) ******
St.5 end St.6 end St.7
0.5
1
1.5
AJ/B
asal
deflection r
atio
****
Stage 5 end
0-45°
45-90°
PA
D
V
St. 7
ROCK inh.
A
0 5 10 15 20 25 300
0.2
0.4
0.6
0.8
1
Deflection (
μm
)
Time (s)
trap on
off
cell interface
De
fle
ctio
n a
mp
litu
de
(μ
m)
C
Time
Experiment Simulation D
experimentmodel
0 5 10����
��
����
0
0.5
1
1.5
Time (s)
Inte
rface d
eflection (
μm
)Experiment
0 5 10Time (s)
Simulation
Time
G
0 1 2 3
00.20.40.60.81
1.2
E
F
B
Distance (cells)
1
Supporting Information
SI Materials and Methods Sample preparation. To mark the cell-cell interface in Drosophila embryos, flies double-labeled with
E-cadherin::GFP (endogenous promoter) and GAP43::mcherry (squash promotor) were selected.
Alternatively, in some cases squash::GFP GAP43::mcherry flies were used. Flies were maintained at
25°C. To obtain embryos, a fresh plate was incubated for 2-2.5 hours. After removing yeast from the
plate, the embryos were washed with 100% bleach for 50 seconds to remove the vitelline membrane.
Embryos at the end of cellularization (stage 5 end) were then selected under a dissection microscope
and aligned on the edge of the coverslip. Alignment was done with the germband visible in the
imaging plane. For experiments with beads and myosin perturbation, embryos were placed in
halocarbon oil and injected using a microinjection setup with either polystyrene beads (1:1000 stock
dilution, Molecular Probes) or with ROCK inhibitor (Y-27632, 10 mM, Invitrogen), respectively.
Embryos were immersed in water for the light sheet microscope setup and in halocarbon oil for
spinning disk imaging.
Optical manipulation and imaging. Optical manipulation of the cell-cell interfaces in individual
embryos was done using a custom-built light sheet microscope (1), coupled with a single beam
gradient trap (1070 nm wavelength, ytterbium fiber laser, IPG Photonics). A 100x water-immersion
lens (1.1 NA, x40, Nikon) was used for imaging as well as for introducing the optical trap in the
imaging plane. Imaging was done using 488 nm and 561 nm excitation lasers. Images were acquired
by an EMCCD camera using a dualview, simultaneous imaging system. Prior to every in vivo
experiment, we calibrated the relationship between galvanometer voltages and laser trap position using
the following procedure: single 500 nm diameter fluorescent polystyrene beads (fluorescence
excitation at 561 nm) were trapped in water and moved slowly by imposing galvanometer voltages
(V1, V2) of the form (V0 cos (ωt), V0 sin (ωt)) with ω < 0.3 rad/s. Images were acquired synchronously
to the voltage commands and successive (X,Y) positions of the bead were localized by a two-
dimensional Gaussian fit. The subpixel localization precision was 25 nm. The measurements were
repeated for different voltage amplitudes (corresponding to trap amplitudes in the image plane < 10
µm). This provided the relationship between (V1, V2) and (X,Y), which was subsequently extrapolated
linearly and inverted to determine (X,Y) as a function of (V1, V2) (Matlab script). This information
was used to provide laser position during every interface deflection experiment. Two kinds of deflections were given to the cell-cell interface: periodic or pull-release. Sinusoidal
oscillations of the galvos were performed to produce linear movements of the laser trap, varying time
periods from 0.3-5 s, amplitudes from 0.3-1.1 µm, and laser power from ~50-300 mW (after objective
lens). For most experiments values were kept constant with a time period 2 s, laser amplitude of
0.5 µm and laser power at 200 mW. For pull-release experiments, the stationary laser trap was
2
switched on at 100 nm – 2 µm from the cell-cell interface for between 10 s - 1 minute. All experiment
recordings were done for galvo-voltage as well as the camera images (either at 561 nm, or both 488
nm and 561 nm excitation).
Quantification of E-cadherin::GFP and Squash::GFP at stage 5 end and stage 7 was done in a Perkin-
Elmer spinning disk microscope using a 100x oil immersion lens.
Quantitative phase imaging uses a transmission light microscope and quadriwave lateral shearing
interferometry, as described in (2).
Data analysis. Kymographs of interface deflections were produced from the movies either in Fiji
(Multiple Kymograph plugin) or using a custom Matlab script. To extract an actual position of the
interface out of the kymograph, a gaussian fit perpendicular to the interface (along the kymograph
line) was performed. At each time step, the peak of the gaussian fit determines the interface position
with subpixel resolution. To determine the localization error, we fixed embryos expressing
Gap43::mCherry and imaged them in the same conditions as in vivo. We then localized cell interfaces
over 100 images, and found that the standard deviation of localization is 35 nm (10 cell interfaces
measured). The interface position, together with the optical trap position recorded through the laser
voltage, was then analyzed in Matlab to measure the response amplitude by maxima detection and
averaging. All fits were performed with Matlab: numerical equations were solved repeatedly,
exploring the space of parameters starting from random values and using the gradient descent method
to minimize error. Statistical analyses were done using the unpaired t-test. For the propagation
analysis, experimental tissue geometries were extracted using the Tissue Analyzer toolbox, courtesy of
B. Aigouy (3), then exported to Matlab to perform the simulations.
The delay between deformations of successive cell-cell contacts in the propagation study was
estimated using a custom Matlab script of “time-sliding fit”. We shifted one signal in time, i.e. we
plotted x(t+∆T) as a function of the trap position xt(t). The time shift ∆T which provides the best linear
fit between x(t+∆T) and xt(t) provides an estimate of the time delay between the two signals. The
confidence intervals were obtained using the nlparci function of the Matlab statistics toolbox.
Model Single junction
The mechanical model for a single junction (Fits in Fig. 1E and 3A) is derived from the constitutive
mechanics of the cortex and from a force balance equation at the interface. The visco-elastic
constitutive equation is given by the so-called standard linear solid (SLS), and relates the horizontal
restoring force to the deflection of the interface:
3
and are elastic parameters (N.m-1), a viscous parameter (m.Pa.s), and the dot denotes a
temporal derivative. At these very low Reynolds number, inertia can be neglected, and the balance of
forces at the interface then simply reads:
where and are the stiffness and position of the optical trap, respectively, and is the damping
coefficient of the interface in the cytosol. The first term on the right-hand side thus corresponds to the
force exerted by the optical trap, while the second corresponds to the viscous drag in the cytosol – and
is therefore proportional to the velocity . This linear system can then be solved for any trap
trajectory ; in particular for our experimental conditions, a sinusoidal oscillation or a pull-release
experiment. Moreover, the relaxation time scales associated to this system can be derived analytically.
Combining the constitutive equation and the force balance in the absence of trap yields:
The solution is in the form / / . In the limit , which we find
is verified from fit values, the two time scales and simplify into:
In that limit, one time scale is related to the viscous component of the cortex, , while the other is
related to the damping coefficient in the cytosol, .
Tissue scale
In the tissue scale simulations, the epithelium is considered as a network of bonds – the cell contacts –
between vertices. Each bond is considered as a visco-elastic segment. The constitutive equation of
each segment, similar to the first equation, is:
where is the tension and is the elongation, . The displacement of each vertex is then
4
computed using the force balance equation between the tension at adjacent contact lines ( ) and damping in the cytosol ( ) (Fig. 3D). The force balance at vertex i thus reads:
i ij This provides direct access to vertices displacements through velocities i. Notably, the midpoint of
the target interface is treated as a 2-way vertex in the simulations. Its movement is imposed as to
mimic the considered experiment. The rest of the vertices move according to the force balance
equation; therefore, their movement ultimately results from the deflection movement of the target
interface. We use fixed (zero displacement) boundary conditions. The areas are not constrained as we
consider small deformations only. For larger deformations, the model would almost certainly require
area or “pressure” constraints.
1. Chardès C, Ménélec P, Bertrand V, Lenne P-F (2014) Setting-up a simple light sheet microscope for intoto imaging of C. elegans development. Journal of visualized experiments: JoVE 87:e51342.
2. Bon P, Maucort G, Wattellier B, Monneret S (2009) Quadriwave lateral shearing interferometry forquantitative phase microscopy of living cells. Opt Express 17:13080–13094.
3. Aigouy B et al. (2010) Cell flow reorients the axis of planar polarity in the wing epithelium of Drosophila.Cell 142:773–786.
5
Supplementary Figure Legends
Figure S1: Optical setup combining light sheet microscopy and optical tweezers. In the light sheet illumination unit, the lasers are mixed by the dichroic mirrors, and enter the AOTF,
which controls the power of each laser independently. Then, the telescope increases the size of the
beam by 5-fold and the periscope brings it to the height of the microscope. The cylindrical lens forms
the light sheet, which is refocused by the illumination objective. The detection unit is integrated in the
upright microscope and is mainly composed by the detection lens, the filter, the tube lens and the
EMCCD. The sample is positioned at the intersection between the illumination and detection paths. A
piezoelectric stage allows vertical (Z) displacements of the sample for 3D acquisition. In the optical
tweezers unit, a near-infrared (NIR) laser beam (1070 nm, continuous wave) is deflected by two
galvanometric mirrors and expanded by a 5-fold telescope. The expanded laser beam is reflected by a
hot dichroic mirror and is tightly focused by the collection objective of the light sheet microscope.
Figure S2: Phase mismatch at cell interfaces. (A) Quantitative Phase Microscopy Image obtained from an early Drosophila embryo (stage 6). The
method uses a transmission light microscope and quadriwave lateral shearing interferometry, as
described in (2). The epithelial cells are observed in cross-section (cartoon, top). Scale bar, 5 µm. The
calibration bar shows the optical path difference in nm. (B) A plot profile along a line (red dotted line
in A) shows that the optical path difference is larger at cell interfaces than inside the cells. Given that
the line defines positions where the geometrical thickness of the embryo is constant, this indicates
there is a refraction index increase at cell interfaces. Arrows of different colors mark the position of 3
interfaces.
Figure S3: Deformation induced by optically-tweezed beads. (A) Position of a 0.46 µm diameter bead (blue line) in the cytosol moving between two trap positions
separated by 0.5 µm (red). From one trap position to another, the bead relaxation is exponential with a
characteristic time given by the ratio of the drag coefficient over the trap stiffness. (B) Snapshots of an
interface deformation induced by a 0.46 µm diameter bead moved by the laser trap against the
interface. The red and green channels correspond to two different positions of the trap separated by
approximately 0.5 µm. (C) Positions of laser trap, bead and interface in an oscillatory experiment with
bead at 100 mW laser power. (D) Positions of laser trap and interface in the same conditions as in (C)
in the absence of bead.
Figure S4: Viscosity measurements obtained from trajectories of individual beads. (A) Image showing 100 nm diameter beads (red) injected in the embryo. Cell contours are labeled by
E-cadherin::GFP. (B) Single particle trajectories superimposed to an image of the cells. (C) Fraction
6
of beads exhibiting diffusive, subdiffusive and superdiffusive behaviors from analysis of the mean-
square displacement (using criteria as described in Reference 1)). Trajectories were acquired at 38 Hz
over a time of 13 to 26 s. Note that the contribution of active fluctuations have been shown to be
important below 10 Hz (Reference 2), and we cannot fully assert that they do not contribute to bead
fluctuations at 38 Hz. Our measurements thus provide only an estimate of the cytosol viscosity
(effective viscosity). (D) Histogram of the viscosity coefficient determined from the analysis of 1348
particles exhibiting free-like diffusion.
Reference 1: Kusumi A, Sako Y, Yamamoto M (1993) Confined lateral diffusion of membrane
receptors as studied by single particle tracking (nanovid microscopy). Effects of calcium-induced
differentiation in cultured epithelial cells. Biophys J 65:2021–2040.
Reference 2: Mizuno D, Tardin C, Schmidt CF, Mackintosh FC (2007) Nonequilibrium mechanics of
active cytoskeletal networks. Science 315:370–373.
Figure S5: Elongation of interfaces adjacent to the optically deformed interface. and denote the length of the optically-tweezed interface, prior to deformation and at maximal
deformation, respectively. Elongation of the interface n (n = 1, 2, 3 or 4), adjacent to the interface 0 is
given by . Left plot shows the ratio of over . The red line is the median, the box
edges are the lower and upper quartiles, and the whiskers display the total range of measurements.
Scale bar, 5 µm.
Figure S6: Relaxation of interfaces after trap release and deformation along the apico-basal axis. (A) After trap release, the interface relaxes with an initial velocity, which is dependent on tension and
onthe damping coefficient in the cytosol . (B) Deformation of a cell interface in regions of the
embryo where the apico-basal axis is in the plane of imaging. The epithelial cells are observed in
cross-section (left panel). Interface prior deflection (top right) and deflected interface (top left). The
deformation extends over a 4-5 µm width along the apico-basal direction (bottom right panel). The
blue lines are eye-guides. Scale bars, 10 µm (left panel), 5 µm (right panels).
Figure S7: Tension normalized to junction length along AP and DV directions.
Figure S8: Pull-release experiments and comparison to simple visco-elastic models. (A) Relaxation dynamics of the interface in linear-log representation. A simple exponential does not
fit the data (cyan solid line). The two characteristic times are visible (purple and green dashed lines as
eye-guides). (B) Alternative visco-elastic models for the pull-release experiments: the Kelvin-Voigt
model (top) is composed of a spring and a dashpot in parallel, while the Maxwell model (bottom) is
composed of a spring and a dashpot in series. In the presence of external viscosity, both predict
exponential relaxation.
7
Figure S9: Mechanical model of the interface and tissue response. (A) Deflection perpendicular to the interfaces is tracked over time along lines perpendicular to cell
interfaces (red). Scale bar, 10 µm. (B) Kymograph of interface deflections in the experimental (left)
and simulated (right) tissues. (C) Deflection of the target (black) and neighbor interfaces (at 1-cell, 2-
cell and 3-cell distance, magenta, green and blue, respectively) in the experimental (left) and simulated
(right) tissues. (D) Spatial decay of interface deflections over the neighboring cells. Comparison
between experiments (black circles) and simulations (red stars). (E) Propagation of deformation for
different values of cytosol viscosity (top: viscosity , middle: , bottom: . 5). A small
viscosity results in a more efficient propagation of the deformation, which becomes limited only by
the fixed boundary conditions. On the opposite, a high viscosity results in a less efficient propagation.
Supplementary Movies Movie 1: Interface deflection produced by a laser trap (red) following a sinusoidal movement of 0.5
µm amplitude and a 2 s time period.
Movie 2: Interface deformation imposed by a laser trap moving a bead against a cell-cell interface.
Movie 3: Propagation of local deformation – comparison between in vivo and in silico experiments.
Filter
Illumination
lens
Tube lens
Dete
ction
lens
Illumination
lens
561 nm
AOTF
Sample chamber
405 nm
488 nm
100x / 1.1 NA
10x / 0.3 NA
Laser beam (front view)
Laser beam (bottom view)
Fluorescence light
Front view
Bottom view
Detection
lens
Cylindrical
lensPeriscope
Beam
expander
Mirror
Dichroic
mirror
Dichroic
mirrorIllumination
Unit
(horizontal)
Detection
Unit
(vertical)EMCCD
Camera
Piezo stage
Z
Dichroic mirror
1070 nm
Beam
expander
Galvo
Galvo
Optical
tweezers
Fig. S1
Position (µm)
Optical path
diffe
rence (
nm
)
nm
QPD
A B
Fig. S2
B
01 2 3 4 5 6 7 8 90
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Time (s)
Positio
n (μm
)
C
Bead
Trap
Interface
Trap
Interface
Time (s)
Positio
n (μm
)
D
01 2 3 4 5 6 7 8 90
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
A
Fig S3
Bead p
ositio
n
in t
he c
yto
sol (μ
m)
-0.2
0
0.2
Time (s)
0 1 2 3 4
BA
0 5 10 15 200
20
40
60
80
100
120
140
160D
Viscosity coefficient (Pa.s)
Fre
quency
C
Number of embryos 26
Number of beads 4650
Fraction of diffusive beads 0.29
Fraction of superdiffusive beads 0.705
Fraction of subdiffusive beads 0.005
Fig S4
Fig S5
0 2 4 6 8 10 12 14 161.95
2
2.05
2.1
2.15
2.2
2.25
Time(s)
V0
Trap switched off
Inte
rface p
ositio
n (μm
)
Fig S6
A
B
Figure S7
St.5 end St.6 end St.7 0
10
20
30
40
50
60
Tensio
n /
junction length
(pN
/μm
)
B
Mem
bra
ne a
nd T
rap P
ositio
ns (μm
)
Fig S8
A
25 26 27 28 29 30 31 32
106
105
104
103
102
101
Time (sec)
Mem
bra
ne P
ositio
n (μm
)
1 2 3
1
Deflection
(μm
)
Distance (cells)0 5 10
0
1
Experiment
0 5 10
0
1
Simulation
A
Time
Experiment Simulation
Time
Time (s)
Inte
rface d
eflection (μm
)
Time (s)
Inte
rface d
eflection (μm
)
B
C D
E
Fig. S9
experimentmodel