Robustness of vulva induction to anatomical variability in C. elegans

(1)

Supervisor:

DR. J.S. VANZON

Examiners: PROF. DR. H.J. BAKKER

PROF. DR. P.R.TEN WOLDE

MSc Physics

Physics of Life and Health

Master Thesis

Robustness of vulva induction to anatomical

variability in C. elegans

by

GUIZELA HUELSZ PRINCE

ID 10395253 August 2014

60 ECTS

September 2013 to August 2014 .

(2)

(3)

Abstract

The induction of the C. elegans vulva has been extensively studied as a model for developmental robustness during the last decades. Despite the many ex-periments that have been performed, fundamental questions regarding how ro-bustness in this system is achieved at the molecular level are still unanswered. Here, we addressed these questions by means of quantitative experiments and mathematical modeling. We found that there is significant anatomical vari-ability among animals during the early stages of vulva induction and that gene expression dynamics show clear differences depending on the initial anatomical configuration. Our approach allowed us to obtain a deeper understanding of the process, in which an interplay between cell induction, lateral inhibition and migration explains how vulva development is such a robust system. Our work might provide a starting point for the integration of quantitative experiments and modeling for the study of the molecular basis of developmental robustness in other systems and organisms.

(4)

(5)

Introduction

1

1.1 Robustness

Living organisms are frequently exposed to internal and external fluctuations and, in order to survive, they must be able to cope with them and continue to be functional. This ability of organismal properties to persist regardless of variations in the conditions they encounter is called robustness. In general terms, it is said that a property is robust with respect to a variable if variation of the former is weakly correlated to variation in the latter.

In biology, robustness occurs on all organizational levels, from protein folding and gene regulation to species persistence and ecological resilience [1]. The study of robustness in the context of developmental biology is of particular interest since a large number of interdependent processes must occur within limited time intervals despite perturbations in order to give rise to a healthy adult capable of reproduction.

For a long time, developmental biologists have qualitatively described nu-merous traits that are in one way or another robust throughout the course of development. For instance, the six legs of insects, the eight legs of spiders, and the seven cervical vertebrae of mammals are known to persist irregard-less of environmental conditions and genetic variations [2]. However, in many cases the molecular mechanisms that lead to such stability are still not clear [3]. Recently, it has become evident that robustness is generally not a result of the action of single genes, but that it rather emerges as a consequence of networks of interacting genes [4]. Thus, a systems-level approach is necessary to understand the origin of robustness. Recent work has identified some gen-eral molecular principles underlying robustness at different biological levels, from biochemical networks in bacteria to cell differentiation in animals [1]. Examples include genetic redundancy in which different genes are able to per-form similar tasks, feedback regulation in which the output of a process serves also as an input, and cooperative biochemical networks in which activation thresholds are generated [5].

The phenomenon of robustness is not only of fundamental importance for cellular and developmental biology, but it also has implications for other bio-logical fields such as evolution and medicine, and even more distant disciplines

(8)

like engineering [6]. In the context of evolution, robustness to mutations allows hidden genetic variations to accumulate from generation to generation [7]. It might happen later that the system is exposed to severe stress which causes the previously hidden variation to be expressed. In some cases, such new traits will be beneficial and will then be transmitted to the next generations. In this manner, robustness to mutations might be an important source of evolutionary innovation.

From the medical perspective, the study of robustness might expand the current approaches to treat various diseases such as cancer [8]. Tumors can be described as robust systems, in which the proliferation of malignant cells is a property that is in some cases able to persist against perturbations brought by anticancer therapies. In this manner, understanding how robust systems work might provide clues about their fragilities and consequently, pave the way to the development of therapies to prevent tumor growth.

In engineering, it is important for systems to be built in ways that prevent fluctuations within a certain range from hindering their performance. Robust man-made systems share many of mechanisms that are used by biological organisms such as redundancy [6]. As a result, the understanding of biological robustness might be helpful for the design and construction of systems that are capable of withstanding severe perturbations.

1.1.1 Sources of variation in development

There are various elements that determine the unique conditions through which each individual develops, such as its genetic background and the external en-vironment [4]. Differences in the genotype among individuals of the same species arise by spontaneous mutations and by recombination which can result in different combinations of previously existing mutations. Regarding the en-vironmental aspect, perturbations are caused by external circumstances such as temperature or availability of nutrients. In addition, even if two organisms are genetically identical and develop in the same environment, internal fac-tors such as stochastic fluctuations in gene expression will prevent them from being exactly alike. Such stochastic factors become specially important when the average number of molecules present in cells is low. Clearly, a difference of one molecule is much more significant when the average is five than when it is 100.

These sources of variability cause organisms to develop by following slightly different paths; yet, these paths eventually converge to reach an invariant outcome. This process is often referred to as developmental canalization.

(9)

Chapter 1 Introduction

Figure 1.1 Two diagrams that have been proposed to illustrate develop-mental robustness. An analogy is made between a ball rolling downhill on a two-dimensional surface and a cell that goes through different de-velopmental stages. The surface consists of a series of bifurcation valleys, which results on a system with a variable number of stable states. In a cell, such stable states correspond to the different phenotypes it can adopt in the process of cellular differentiation. A Landscape proposed by Waddington [9] representing a cell that starts on a monostable undif-ferentiated state. Through the appearance of pitchfork bifurcations Ψ, the cell continues through one of two mutually exclusive paths. B A sec-ond landscape proposed by Ferrel [10] depicts a scenario where a cell can originally be found on different stable states, but through the appearance of saddle node bifurcations SN is induced or canalized towards a single stable differentiated state. (Adapted from [10]).

1.1.2 A physical view of developmental canalization

Some abstract physics-inspired ideas have been proposed to illustrate devel-opmental robustness and canalization. In 1957, Waddington introduced what he called the epigenetic landscape [9] which is shown in Figure1.1A. In this diagram, the ball represents a cell and each valley is a possible cell fate that it can adopt. Therefore, as the ball rolls down through the landscape, it is canalized into one of many possible discrete differentiated states, without the possibility of adopting any intermediate fate. This drawing is of course an oversimplification of the process since only one variable along the horizontal axis is employed to describe the whole set of phenotypes a cell can have. In re-ality, such a surface would correspond to a hypersurface in a multidimensional space.

Waddintong’s landscape, however, only considers a single starting point. This means that the system starts in a monostable state which, through the appearance of a series of bifurcating valleys (see Figure1.1A), eventually results in a multistable system. According to Ferrel [10], this diagram is suitable to describe cell differentiation through lateral inhibition. In this scenario, a cell starts in a single undifferentiated state and through competition with other

(10)

cells it can adopt one of different mutually exclusive phenotypes. Therefore, this process is analogous to a single valley that splits into two from which the ball can roll down into only one.

Ferrel proposes that for other processes, such as cell fate induction, a land-scape that consists of disappearing valleys (see Figure1.1B) provides a more accurate description. During cell induction, one cell usually generates an in-ductive signal that is received by another cell or group of cells, and indicates them to adopt a particular fate. Consequently, the receiving cells might orig-inally stand on a variety of stable states but, upon receiving the inductive signal, there is only a single fate that they are able to adopt. The analogous landscape is then one in which the ball is initially situated in one of many valleys, and all except one will eventually disappear, as shown in Figure1.1B. These two surfaces provide a simple yet illustrative way of understanding canalization. Still, they remain a hypothetical concept since concrete mech-anisms underlying canalized developmental processes have not been found. Finding and describing one of such mechanisms was a major motivation for the work that will be presented in this thesis.

1.1.3 Approaches to studying developmental robustness

The predominant approach to the study of robustness is the experimental introduction of perturbations to biological organisms, such as the nematode Caenorhabditis elegans, the fruit fly Drosophila melanogaster, and the mouse Mus musculus. For decades, biologists have tested the effects of environmental perturbations such as heat shock or exposure to different chemical environ-ments. Genetic perturbations have also been introduced, whether they origi-nated from a random generated mutation or by genetic engineering techniques. After a certain perturbation is applied, robustness is quantified according to the resulting variation of the trait or traits of interest [4]. However, most of the studies have focused on the average developmental outcome, ignoring the dynamics of the process and the variability among individuals.

With the advent of powerful computers, mathematical modeling and sim-ulations have proven to be essential techniques when it comes to the study of systems biology. Biological networks are usually conformed by a large amount of components which interact with one another in complex non-linear ways [11]. Simulations are capable of probing large sets of parameters which can be difficult or not possible at all to introduce experimentally. Therefore, modeling might provide predictions that can be later be confirmed by experimentation, as well as clues about which are the key parameters that play a role in the system of interest. Extensive theoretical work has been performed to address developmental robustness, but most of the current experimental results lack

(11)

Chapter 1 Introduction

quantitative measurements at the molecular level that could serve to constrain the existing models [4].

In this work, we will study a robust system with the use of both experi-mental perturbations and modeling. We will focus on the development of the vulva in C. elegans and use quantitative experimental techniques to study the dynamics of the process. We will also develop a simple mathematical model that can reproduce the essential components of our experimental results, with the aim of further understanding the key elements that provide robustness to the system. The next section will present an introduction to C. elegans and its relevance in developmental biology, as well as a detailed description of the process of vulva development.

1.2 C. elegans

The nematode Caenorhabditis elegans, a 1.5mm-long roundworm (Figure1.2), has become an important model organisms for the study of many disciplines within biology. Experimental work with such organisms presents many ad-vantages: they reach adulthood within three days, they are small and easy to maintain, they produce around 300 offspring, and they are transparent, just to name a few. Research with C. elegans has led to numerous advances in biology and medicine. Some discoveries that have been made in C. elegans include cel-lular mechanisms that regulate apoptosis, the existence of RNA interference, mechanisms of axon guidance, mitochondrial involvement in oxidative stress and aging, and signal transduction pathways, among others [12]. Despite the evident anatomical differences between these nematodes and higher species, it has been discovered that many genes present in the worm’s genome have corresponding homologs in humans and other animals [13].

Figure 1.2 Adult hermaphrodite C. elegans and two eggs (Image from [14]).

Many C. elegans wild-type strains have been isolated from different places around the world. However, most of the research is done using the strain termed N2, since it is the line that was introduced as a model organism in biology by Sydney Brenner in 1964 [13].

(12)

C. elegans can be either self-fertilizing hermaphrodites or males. Males are not found frequently in populations under standard conditions, but they appear more frequently when the animals are faced with stressful environments [14]. The ability of hermaphrodites to self-fertilize and the possibility of in-ducing the presence of males through heat shock treatment allows the crossing of different mutant strains, making it possible to combine multiple mutations in a single animal. For this reason, C. elegans is a very useful organism for genetic assays.

C. elegans is a particularly interesting specimen for the study of devel-opmental robustness since, with very few exceptions, every worm develops by following identical patterns of cell divisions and differentiation. All adult hermaphrodites contain 959 somatic cells and the complete cell lineage of each has been characterized. This task was facilitated by the fact that these ani-mals are transparent and cells can be easily distinguished using Nomarski or Differential Interference Contrast (DIC) microscopy [14].

1.2.1 Life cycle of C. elegans

The life cycle of the hermaphrodite animals consists of an embryonic stage, four larval stages named L1 to L4, and adulthood. The duration of each stage varies according to the environmental temperature; faster development occurs at high temperatures (∼25°C) while it is slower at low temperatures (∼15°C) [14].

During the embryonic period, C. elegans develop inside an egg and hatch at about nine hours after fertilization. Different developmental processes which include cell differentiation, division and migration occur during the larval stages, which in total last about 38 hours. Finally, the animals reach adulthood and are able to start laying eggs eight hours later.

1.2.2 Vulva development as a model for robustness

The vulva of C. elegans hermaphrodites is the organ that connects the uterus with the outside environment; it is located in the ventral side of the mid-body and it allows egg-laying as well as mating. This organ has become one of the best studied examples of organogenesis since it is a simple and robust system that involves the interaction of various signaling pathways [15].

The process of vulva development involves a somatic gonadal cell called the anchor cell (to be called AC from now on), and a row of six ventral cells called vulval precursor cells (to be called VPCs from now on), meaning that they all have the potential of later becoming vulval cells. Each VPC is termed Pn.p, where n stands for the numbers 3 to 8, and in general, the AC is positioned

(13)

Chapter 1 Introduction P8.p P7.p P6.p P5.p P4.p P3.p LIN-3 VPCs

}

Figure 1.3 Representation of the vulva induction process. The anchor cell (AC) secretes LIN-3 which forms a gradient and reaches basal side of the vulval precursor cells (VPCs). Upon receiving LIN-3, signal transduction pathways are activated in the VPCs resulting on each cell adopting a particular cell fate. The cell that is closest to the AC, P6.p, becomes a primary cell, while its neighbors, P5.p and P7.p adopt a secondary fate. These three cells will eventually form the worm’s vulva while the rest will become part of the skin.

closer to P6.p. Vulva induction starts when the AC secretes the protein LIN-3 which presumably diffuses away and forms a gradient. When VPCs receive this signal, it induces them to adopt one of three cell fates, commonly referred to as 1° (primary), 2° (secondary), or 3° (tertiary) fates. The VPC that receives the largest inductive input, usually P6.p, will become a 1° cell, the neighboring VPCs, P5.p and P7.p, will become 2° cells and the rest will adopt a 3° fate (Figure 1.3). VPCs that become 1° and 2° cells will give rise to the vulva during L4 stage, while those that adopt a 3° will become part of the epidermis. All VPCs have the capability of responding to the AC signal, thus the 2° 1° 2° pattern can be centered on other cell than P6.p, although this is rare in wild-type conditions. A shift in this pattern will not prevent the formation of a functional as long as the 1° cell is flanked by two 2° cells. In wild-type conditions, the formation of the 2° 1° 2° pattern is extremely robust; mistakes only occur in about 0.25% of animals [16].

The process of induction involves the interaction of two signaling pathways that are known to be highly conserved in animals [17, 18]. The AC communi-cates with the VPCs via the so called Ras pathway whose activation promotes 1° cell fate, while the VPCs communicate with its two lateral neighbors via the Notch signaling pathway whose activation induces 2° fate. A simplified illustration describing how both pathways interact with each other within and between two VPCs is shown in figure 1.4A, where positive regulation is rep-resented by pink arrows, while negative or inhibitory regulation is depicted in green. In summary, when a VPC receives the inductive signal, the Ras path-way is activated promoting the production of 1° cell fate markers while sending a Notch signal to its neighbors. In turn, when a VPC receives a Notch signal, it decreases its response to the inductive signal and it is induced to become a 2° cell.

(14)

A

Inductive signal

B

from AC Ras signaling Notch signa-ling Ras signaling Notch signa-ling 1º fate 2º fate 1º fate 2º fate VPC 1 VPC 2 LIN-3 LET-23 LAG-2 LIN-12 APX-1 lst-1,2,3,4 ark-1 LAG-1 lip-1 mir-61 VAV-1 LET-60 MEK-2 LIN-45 egl-17 LIN-1 lag-2 apx-1

Ras pathway Notch pathway

MPK-1 Ligands Receptors Signal transduction Transcription factors Target genes

Figure 1.4 A Summary of the pathway interactions within and between two VPCs. Positive interactions are depicted in pink and negative interac-tions are shown in green. B Overview of the main proteins, genes and interactions that play a role in the Ras and Notch signaling pathways within one VPC [15].

An overview of the most important components as well as protein-protein interactions and the target genes of both signaling pathways in one VPC is presented in Figure 1.4B. Throughout this thesis, we will often refer to LIN-3 (the inductive signal), LET-23 (the VPC receptor of LIN-3), LIN-12 (the Notch receptor in VPCs), and apx-1 and lag-2 (Notch ligands whose production is activated upon the reception of the inductive signal and activate Notch receptors in adjacent VPCs).

After more than three decades of research, there is still a question that has no clear answer, namely if lateral inhibition by Notch signaling is an essential component that allows the system to be so robust. Some experiments support the hypothesis that vulva induction is robust against perturbations without the use of lateral inhibition and have led to the proposal of the sequential induction model (Figure 1.5A). In contrast, other experiments suggest that Notch signaling is essential for the formation of the 2° 1° 2° pattern by prevent-ing adjacent VPCs from adoptprevent-ing 1° cell fates which has led to the graded induction model (Figure1.5B).

The graded model is supported by the fact that in a strain with a mu-tation in lin-12 (the gene that encodes for the Notch receptor LIN-12) the observed phenotype is the appearance of adjacent 1°cells [19]. However, Notch signaling is also involved in the specification of the AC, causing most animals to develop two ACs. Therefore, it is likely that the dosage of LIN-3 on this mutant is higher than normal, which leads to unclear evidence as to whether the inhibitory role of Notch signaling is significant in wild-type conditions. On

(15)

Chapter 1 Introduction P5.p P6.p P7.p Sequential induction Notch ligand Notch receptor P5.p P6.p P7.p Graded induction

A

B

Figure 1.5 A Model in which the role of Notch signaling is to induce sec-ondary cell fate and its inhibitory properties are not necessary for a robust induction. Initially, only P6.p receives enough LIN-3 to activate the pro-duction of Notch ligands. Later, LIN-12 receptors on the adjacent cells are activated leading to the adoption of secondary cell fate (green arrows). Such process results in the 2°1°2°pattern. B Model in which Notch sig-naling is an important component for the prevention of adjacent 1° cells in addition to inducing 2° fate. At the beginning of the process, LIN-3 reaches the closest cells to the AC. Upon the activation of Ras signal-ing, cells start producing Notch ligands according to the amount of signal received. The cell that produces the highest amount of Notch ligands causes the adjacent cells to stop responding to LIN-3 (pink lines) and promotes their adoption of secondary fate (green arrows), resulting in a 2°1°2°pattern.

the other hand, recent evidence supporting the sequential model concerns the loss of 2° fates in P5.p and P7.p rather than the appearance of adjacent 1° cells in animals where lin-12 was downregulated in a tissue specific manner [20]. In these experiments, there was only one AC present suggesting that the LIN-3 levels were close to wild-type.

So far, vulva induction has been studied by considering average outcomes at the cell fate level, while neglecting the dynamics of the process and the variability among animals. Moreover, after so much time and effort dedicated to the study of this system, the issue has always been addressed with the assumption that the VPCs and the AC are positioned correctly during the entire process. That is, they all consider that P6.p is always directly adjacent to the AC. However, it is probable that the extensive amount of developmen-tal processes that control growth and cell positioning might result in slight

(16)

anatomical differences among animals. If that is the case, vulva induction must not only be robust to environmental and genetic changes, but also to stochastic processes that lead to variable cell positions. Understanding how the system responds and copes with this type of perturbations might provide some insight about the actual role that Notch signaling plays to achieve a high degree of robustness, i.e. if its Ras-inhibitory properties are essential for a 1° versus non-1° decision or not.

The purposes of the research study presented in this thesis are to quantify the anatomical variability in wild-type animals to determine if it is indeed a significant phenotype, to understand how vulva induction is robust against such perturbation, and to determine if the lateral inhibition of Notch signaling is a crucial component for the network to be so robust.

(17)

Previous work

2

With the aim of obtaining a quantitative insight into the dynamics of gene expression patterns during the process of vulva induction, J.S. van Zon et al. [21] performed a series of experiments in which a particular focus was made on the Notch ligands apx-1 and lag-2. Such genes were chosen since they can be used as markers for 1° cell fate and at the same time they act as the agents of lateral inhibition between VPCs. Thus, they could provide an insight into the workings of the two pathways.

The expression of Notch ligands was quantified with the use of single-molecule fluorescence in situ hybridization (smFISH) which allows the visual-ization of single mRNA molecules (the technique is explained in further detail in section3.2). Combining the experimental results with mathematical model-ing, the study revealed previously unknown aspects of the way in which VPCs modulate the signal received by the AC. The next sections will provide an overview of the results that are directly relevant for the research performed in this thesis, and the way in which they led to questions regarding robustness to anatomical variability in the AC position with respect to the VPCs.

2.1 Notch ligand expression during vulva

induc-tion

Single-molecule FISH is a very sensitive tool with which it is possible to mea-sure the amount of mRNAs in a cell with high precision. However, this tech-nique does not allow for time-lapse imaging since specimens have to be fixed for observation. Therefore, dynamic behaviors can only be obtained by imaging many animals, where each individual provides information about the expres-sion levels at the moment of fixation. A problem now arises since it is then important to determine the exact developmental stage at which the animal was during fixation. This issue was addressed by the use of an anatomical marker in order to assign an absolute developmental time to each fixed animal.

(18)

A B

Figure 2.1 A C. elegans carrying a lag-2p::GFP transgene imaged at different time points throughout development. Yellow markers represent DTCs and the dashed line shows the gonad axis. The scale bar corresponds to 20µm. B Gonad length as a function of time measured with respect to the start of L3. Each color corresponds to one of the five observed animals. The gray area represents the L2 lethargus, a period of inactivity prior to the transition to a new stage, and the black line is the best fit to the data obtained for all animals (see equation2.1). (Images from [21])

2.1.1 Gonad length as a proxy for developmental stage

The gonad of C. elegans grows along the long axis of the animals and it is known that its length increases during the L2 to L4 stages in a stereotypical manner [22]. This process occurs as germ cells proliferate following a signal provided by two cells called distal tip cells (DTCs). The DTCs lie at either end of the gonad and the distance between them provides a good measurement for gonad length. Since DTCs express lag-2 [23], one of the targeted Notch ligands in the study, the lag-2 smHISH signal could be used to identify the DTCs in fixed animals.

To obtain the relationship between gonad length and time, live animals containing a lag-2p::GFP transgene were imaged throughout the course of the L2, L3 and mid-L4 developmental stages. This transgene allows the animals to produce the green fluorescent protein (GFP) in a way that is regulated by the lag-2 promoter (lag-2p), a DNA sequence that dictates controls the expression of the gene. As a result, DTCs can be visualized and the distance between them can be measured as the gonad grows along the animal (Figure 2.1A).

The results were remarkably similar from animal to animal, and the data were fitted to a piecewise polynomial given by

G(t) = (

17.7 + 0.54t for t < 0,

17.7 + 0.54t + 0.72t2 for t _{≥ 0,} (2.1) where G is the gonad length in µm and t is time in units of hours with t = 0

(19)

Chapter 2 Previous work 0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 120 lag-2 mR NA #

Hours after start of L3

apx -1 mRNA # 0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 120

Hours after start of L3 P7.p

P6.p P5.p

Figure 2.2 Number of apx-1 and lag-2 mRNAs in the three closest cells to the AC as a function of time after the start of L3.

corresponding to the start of L3 (Figure 2.1B).

Once the correlation between gonad length and developmental stage was determined, the next step was to quantify the amount of mRNAs in each VPC in a population of animals that were fixed at different time points during L3.

2.1.2 Expression dynamics of Notch ligands

The targeted mRNAs to study expression dynamics were the Notch ligands apx-1 and lag-2. The resulting mRNA levels in P5.p, P6.p and P7.p are shown in Figure2.2, where each point corresponds to a single animal. The considered range of time, from 0 to 12 hours after the start of L3, corresponded to the interval between the specification of the AC, which occurs around the L2 to L3 transition, and the first VPC divisions that occur upon the L3 to L4 transition. It was observed that Notch ligand expression had different behaviors during early and late induction. Initially (_{∼0-4 hours), apx-1 is expressed at low levels} in P5.p and P7.p and at slightly higher numbers in P6.p. Additionally, lag-2 is not present in P5.p and P7.p, but low levels are found in P6.p. In contrast, after this period the number of both ligands increased steeply in P6.p, while it vanished in most P5.p and P7.p cells. Furthermore, the final levels of lag-2 were approximately twice as much as the levels of apx-1. The increase in expression was found not to be due to the accumulation of mRNAs over time, since production and degradation rates were in a steady state relative to the timescale of vulva induction. Therefore, Notch mRNA levels were considered hereafter as an instantaneous readout of Ras signaling levels in a VPC.

By using a lin-12 mutant, they prevented lateral inhibition between VPCs with the aim of measuring the strength of Ras signaling in cells located at different distances from the AC. The studies revealed that the levels of Ras signaling in VPCs decayed as a function of the distance to the AC with a half-width of _{∼15 to 20 µm. The question of anatomical variability arose as}

(20)

A

B

AC P5.p P7.p P6.p 10µm 11 hours after start of L3

10µm 3 hours after start of L3

P5.p P6.p P7.p

P4.p

AC

10µm 2 hours after start of L3

P5.p P6.p _P7.p

AC

Figure 2.3 A Two early L3 wild-type animals where the AC is correctly po-sitioned and severely misplaced respectively. B Late L3 wild-type animal with a correctly positioned AC. Blue corresponds to cell nuclei stained with DAPI.

a spin-off of these findings since this decay distance is of a similar magnitude as the length of a VPC along the long axis of the animals (_{∼15 to 17 µm).} Consequently, variations in the AC position with respect to the VPCs might result in two cells receiving a similar dose of inductive signal. If such anatomi-cal variability occurs in wild-type animals, the system must have a mechanism to prevent the induction of adjacent 1° cells since this phenotype is rarely ob-served in such animals. It follows then that the first question to be addressed should regard the existence of a significant amount of wild-type animals with an initial incorrect cell configuration.

2.2 Anatomical variability in wild-type animals

When cell positions on fixed wild-type animals were analyzed, it was noted that the AC location with respect to the VPCs was highly variable among young L3 animals, while older ones presented an AC centered on P6.p. Examples are shown on Figure 2.3, were panel A corresponds to two young L3 animals, one with a correctly placed AC and another with a severely misplaced AC located between P5.p and P6.p, and panelBrepresents a late L3 animal with a correct cell configuration.

(21)

Chapter 2 Previous work

A

C

AC correctly placed AC mildly misplaced AC severely misplaced

B

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.1 0.2 0.3 0.4 0.5 R

Figure 2.4 A Definition of the relative AC position R given by the distance from the closest VPC to the AC normalized by the distance between the two closest VPCs. B Color code used the categorize animals according to their degree of AC displacement. C Relative AC positions obtained for N2 animals throughout the course of L3. Each point corresponds to a single fixed animal. The black line represents a sliding average with a window size of two hours and evaluated every 0.5 hours. The gray area represents the average _{± the standard deviation. Arrows indicate the} animals pictured in Figure2.3.

position, was defined as

R = |xAC− xc| |xsc − xc|

, (2.2)

where xAC is the position of the AC, xcis the position of the closest VPC to the AC, and xsc is the position of the second closest VPC to the AC (see Figure

2.4A). Thus, R varies from zero, when the AC is centered on the closest VPC, to 1_/_{2, when the AC is located exactly in between two VPCs. To categorize} the animals according to the degree of AC misplacement, the range of R was divided in three intervals were each is color-coded as shown in Figure 2.4B. Red represents animals in which the AC is correctly positioned, green are those in which it is mildly displaced, and in blue are the ones in which it is severely misplaced.

The analysis of several N2 animals yielded the results shown in Figure2.4C, with each point representing one fixed animal. During the first half of vulva induction (0-6 hours), 23.5% of the animals presented a severe displacement (blue points), while this number surprisingly decreased to 2.5% during the second half (6-12 hours). When calculating a sliding average of the data (black

(22)

line), it is clear that the average relative AC position decreases over time and most late L3 animals have a properly positioned AC. Additionally, the standard deviation (gray area) during early L3 (_{∼2 hours) is about 2.5 times larger than} during late induction (_{∼11 hours). These results suggest that there is likely} an active mechanism that corrects for the initial AC misplacement.

2.3 Remaining questions and approach

The main questions that will be addressed in this thesis concern the mechanism that leads VPCs to acquire the correct cell fate pattern and prevent the in-duction of adjacent 1° cells considering that the initial cell positions are highly variable, as well as the mechanism that directs cells to eventually rearrange and correct the initial AC misplacement.

As mentioned earlier, the percentage of wild-type animals with severely misplaced ACs is 23.5% during early induction. Since our aim is to determine how the system is able to cope with such initial positions, the analysis of a large number of animals is required so we sought for a mutant strain in which this percentage could be higher. We therefore turned to a mutant in which the gonad, and thus the AC, is shifted with respect to wild-type animals. In this strain, we quantified the dynamics of Notch ligand expression as well as the AC positions throughout L3. With our findings, we were able to generate a hypothesis which we later tested by perturbing the Notch and Ras signaling pathways with the use of mutant strains, and by the developing of a simple mathematical model that addresses the current knowledge of both pathways.

(23)

Techniques and methods

3

3.1 Handling and maintenance of C. elegans

In the laboratory, C. elegans are kept on Petri dishes containing nematode growth medium (NGM) agar. This medium is used to allow the growth of the OP50 strain of Escherichia coli which serves as a food source for the animals. These bacteria are not able to synthesize uracil, an essential nutrient for their survival, and thus can only grow forming a thin layer over the surface of the nutritious NGM agar. In these conditions, animals can be maintained at temperatures ranging from 15°C to 25°C depending on the research interest. For the experiments carried out during this project, animals were kept at 20°C except when otherwise noted.

To study animals at a particular developmental stage, it is important to generate synchronized populations of animals. One way to synchronize the animals consists of adding water to a Petri dish containing many eggs. These tend to stick to the agar surface while the larvae and adults are aspirated and discarded. The remaining eggs are allowed to hatch for a specific amount of time, after which larvae are transferred to new plates with plenty of food available. A different way of synchronizing consists on suspending many gravid hermaphrodites on an active bleach solution which results on the disintegration of the adult animals while the unlaid eggs are protected by their shell. Finally, the eggs are deposited on fresh plates seeded with bacteria. In both cases, the the process results in populations of animals progressing through development within a narrow time range. The bleaching technique is particularly useful when studying strains that have the tendency of being egg-laying defective.

The above descriptions represent only the most important aspects of C. elegans maintenance, more detailed protocols can be found in reference [24]. The C. elegans strains that were used during this research project are listed in appendixA.

3.2 Single-molecule FISH

Single-molecule fluorescence in situ hybridization (smFISH) is a novel tech-nique that allows the visualization and quantification of individual mRNA molecules [25]. The technique consists on the use of a library of short DNA

(24)

}

Fluorescently-labeled DNA probes Target mRNA

}

Hybridization

}

Figure 3.1 Illustration of the smFISH technique. A library of many short fluorescently-labeled DNA probes are complementary to different regions of an mRNA. Each probe is not bright enough to be detected, but when multiple probes hybridize to the target mRNA the signal becomes bright enough to be detected as a diffraction limited spot. (Image provided by J.S. van Zon)

probes that are complementary to different regions of the same target mRNA, as illustrated in Figure 3.1. Each probe is coupled to one fluorophore which by itself is not visible, but only when numerous probes bind or hybridize to the same transcript, they become visible as a diffraction-limited spot under the microscope. In this manner, non-specific binding of probes is not likely to contribute to the detected signal, reducing considerably the background noise. One of the advantages of smFISH as a method for quantifying gene ex-pression is that the spatial exex-pression patterns within different cells in a tissue can be determined with a high degree of sensitivity, in contrast to other tech-niques in which either sensitivity or spatial resolution have to be sacrificed in order to increase the other. Additionally, smFISH probes are able to efficiently penetrate through tissues due to their small size, allowing a precise detection of very low numbers of mRNAs.

The downside of smFISH is that temporal dynamics on a single sample cannot be obtained, since the tissues need to be fixed before testing. However, one can circumvent this issue by analyzing many samples, each at a particular time point during the process of interest, from which the dynamics can later be inferred.

The next sections will provide an overview of the most relevant steps for the use of smFISH in the study of C. elegans development and further details can be found in reference [26]. Step descriptions are focused on the procedures used in this project.

(25)

Chapter 3 Techniques and methods

3.2.1 Sample preparation

The initial step consists on the synchronization of animals, either by bleaching or washing, depending on the strain and timing preferences. Nematodes are allowed to grow in the desired conditions, usually at 20°C and with plenty of food, until they reach the developmental stage of interest. Then, animals are collected and fixed by suspension in a 4% formaldehyde solution. This treatment results in the formation of covalent bonds that cross-link different proteins. These bonds presumably preserve the structural properties that the animal had at the moment of fixation. Later, animals are suspended on an ethanol solution with the aim of disrupting the lipidic cell membranes, thus making the tissues permeable to the smFISH DNA probes.

Once the animals are permeabilized, they are ready for hybridization. Fixed animals are suspended in a hybridization solution containing a small concentration of the desired smFISH probes. Additionally, the solution con-tains other components to improve the hybridization efficiency, including dex-tran sulfate and formamide. The former is a compound that is only permeable to small molecules such as water. This results in an effective decrease of probe concentration, since the volume occupied by dextran is not available for the larger DNA probes. The effect of formamide is to reduce the effective temper-ature at which the hydrogen bonds of double stranded DNA or RNA molecules are broken causing both strands to separate. When the formamide concentra-tion is in within a certain range, weak unspecific binding is prevented, while correctly bound probes are stronger and remain intact.

For the purposes of this research project two smFISH probes were used targeting the Notch ligands apx-1 and lag-2. Probes for the former gene were coupled to the fluorophore Cy5 which absorbs maximally at 649nm (red) and emits at 670nm (far red). Probes for the latter gene were coupled to Alexa 594 which has an absorption peak at 590nm (yellow) and an emission maximum at 617nm (red).

Animals are left in the hybridization solution overnight and are then sus-pended on a solution containing 4’,6-diamidino-2-phenylindole (DAPI). DAPI is a molecule that becomes highly fluorescent when it binds to adenine-thymine regions in DNA; it is consequently used extensively as a marker for cell nuclei. DAPI has an absorption peak at around 360nm (UV) and an emission maxi-mum close to 460nm (blue).

3.2.2 Imaging

Images are obtained with a standard wide-field fluorescence microscope with an oil immersion 100x objective. For each animal, 30 images per channel are taken at different z-positions, spanning 10.5µm with a step size of 0.35µm and

(26)

A

B

C

D

Figure 3.2 Examples of images obtained with smFISH. The pictured animal is a lin-1 mutant in which Notch ligands are constitutively expressed in VPCs. lag-2 and apx-1 mRNAs are shown in red and green respectively. A and C correspond to the raw images taken on the Alexa594 (lag-2 ) and Cy5 (apx-1 ) channels. B and D are the same images after the LoG filtering process. In these, the background is greatly reduced and the contrast between smFISH spots and the background is amplified.

centered in the plane of focus corresponding to the midsection of the animal.

3.2.3 Image analysis

Image analysis to obtain the number of mRNA molecules in the regions of interest is usually done with the use of custom-written software. During this research project, I employed a MATLAB script that was previously written by J.S. van Zon. The procedure starts with manually defining the perimeters of the cells that will be studied. The image is then convolved by a Laplacian of Gaussian (LoG) function, which is used to describe diffraction limited spots, in order to sharpen the smFISH spots. Figure3.2shows an example of the results obtained after the filter is applied. PanelsAandCshow the raw images for the Alexa594 and Cy5 channels respectively. Panels B and D represent the same images after the filtering process, where only spots resembling LoG functions remain. To eliminate the remaining background spots, a threshold is then manually set so that only clearly visible smFISH spots are left. Afterwards, connected regions are determined and local maxima within each of these are counted to yield the number of mRNAs present in the selected cells.

(27)

Experimental results and

discussion

4

4.1 Increasing anatomical variability

To increase the percentage of animals that have severely and mildly displaced ACs upon the start of L3 we turned to the dig-1 strain which exhibits defects in gonad placement. In most of these animals, the gonad is shifted anteriorly (towards the head), in such a way that a fully functional vulva still develops but is often centered in P5.p rather than P6.p. Around 10% of animals have a dorsally (towards the back) positioned gonad which prevents the vulva from developing correctly. Such animals were not considered in this study due to their severe deviation from wild-type conditions.

The dig-1 gene encodes a protein that has been suggested to have an ad-hesive function important for the positioning of the gonad primordium during embryogenesis [27], which is consistent with the gonadal displacement observed in larvae and adults. So far, there is no evidence that the impaired function of DIG-1 directly affects the process of vulva induction, which is supported by the fact that, except for those with a dorsal gonad, animals are able to lay eggs and mate properly.

To study how the induction process varies as a function of the AC position, we proceeded by quantifying the relative AC position and expression of Notch ligands throughout L3. We will discuss first our findings regarding variability in the AC position and we will turn later to describe the gene expression dynamics that we observed.

4.1.1 Anatomical variability in dig-1 mutant

When analyzing the AC and VPC positions, the percentage of young L3 dig-1 animals with a severely misplaced AC was increased when compared to N2. The acquired data is shown in figure 4.1A, which is analogous to figure 2.4A

obtained for N2. Similarly to what was observed for the wild-type strain, this fraction became lower at later stages. During the first half of the induction process (0-6 hours), 40.5% of the worms had a severe AC misplacement, while this fraction decreased to 18% in the second half (6-12 hours). In addition, similarly to the wild-type strain, the standard deviation was about two times higher during early L3 (_{∼2 hours) than in later induction (∼11 hours). For}

(28)

a better visualization of the data, figure 4.1B shows the fraction of worms belonging to each category of AC misplacement for subsequent one hour time intervals. The plot shows that there is a tendency for the percentage of worms with correct configurations to increase with time.

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.1 0.2 0.3 0.4 0.5 R

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.2 0.4 0.6 0.8 1 Fr acti o n of worms

Hours after start of L3 A

B

Figure 4.1 A Relative AC positions obtained for dig-1 animals throughout the course of L3. Arrows indicate the animals pictured in figure4.5. The black line represents a sliding average with a window size of two hours and evaluated every 0.5 hours. The gray area represents the average _{± the} standard deviation. B Fraction of animals that belong to each category of AC displacement per time intervals of one hour. Colors indicate the different categories of AC misplacement in panel A.

When comparing the sliding averages _{± the standard error obtained for} the relative AC position in both N2 and dig-1, as shown in figure 4.2, it is remarkable that even though the latter strain starts at and keeps having a high degree of AC misplacement, by the end of induction they both arrive at a similarly low degree of misplacement (R_{∼ 0.1) after 11 hours.}

In the N2 curve, a steep decrease occurs between 5 and 6 hours; in contrast, the dig-1 curve decays gradually with time. Such difference could probably be due to the lower number of N2 animals that were analyzed (73) compared to amount that were evaluated for dig-1 (247). The data obtained for N2 are not sufficient to draw conclusions about the abruptness of the transition, which prevented us from determining its significance.

(29)

Chapter 4 Experimental results and discussion 0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.1 0.2 0.3 0.4 0.5 N2 dig-1

R

Figure 4.2 Average relative AC position as a function of time for N2 (pink) and dig-1 (blue) animals. Lines represent sliding averages with a win-dow size of two hours and evaluated every 0.5 hours. The shaded areas represent the average_{± the standard error.}

4.1.2 Notch ligand expression dynamics in dig-1 mutant

To study Notch ligand expression dynamics in dig-1 animals, we first con-sidered that the vulva of these mutants is often centered on a different VPC than P6.p as in the N2 strain. This could lead to different behaviors, since it is thought that, even though all VPCs are capable of responding to the inductive signal, their ability of transducing the signal is not equivalent [15]. Therefore, we first checked whether in dig-1 animals the dynamics of expression in the VPCs that showed the highest mRNA numbers present significant differences when compared to wild-type animals.

Figure4.3 shows a comparison between N2 and dig-1 animals of the apx-1

lag -2 mRNAs 0 40 80 120 140 100 60 20

0 1 2 3 4 5 6 7 8 9 10 11 12 Max 0 1 2 3 4 5 6 7 8 9 10 11 apx-1 mRNAs 0 40 80 120 140 100 60 20 Max 12 lag -2 mRNAs 0 40 80 120 140 100 60 20 Averag e max apx-1 mRNAs 0 40 80 120 140 100 60 20 Averag e max

0 1 2 3 4 5 6 7 8 9 10 11 12 0 1 2 3 4 5 6 7 8 9 10 11 12

N2 dig-1

A B

Figure 4.3 Comparison of apx-1 and lag-1 expression levels in the VPC that showed the highest amount of mRNAs in N2 (pink) and dig-1 (blue) animals. A mRNA numbers for each animal that was analyzed. B Sliding averages of expression levels calculated with a window size of two hours and evaluated every 0.5 hours. The shaded areas represent the average_± the standard error.

(30)

0 1 2 3 4 5 6 7 8 9 10 11 12 −1 −0.5 0 0.5 1 apx − 1 0 1 2 3 4 5 6 7 8 9 10 11 12 −1 −0.5 0 0.5 1 lag − 2

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.2 0.4 0.6 0.8 1 Fracti on of worms 0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.2 0.4 0.6 0.8 1 Fracti on of worms

Expression in closest cell

Expression both cells

Expression in second closest cell

A

B

Q

C

Figure 4.4 A Color codes used to categorize animals according to their rel-ative ligand expression. B Relrel-ative ligand expression of apx-1 and lag-2 in dig-1 animals throughout the course of L3. Each point corresponds to one worm and is colored according to the AC displacement. The green, pink and purple bands in the background show how the animals are cate-gorized according to the relative ligand expression. C Fraction of animals that belong on each category of relative expression per time intervals of one hour.

and lag-2 mRNA numbers. Panel A presents an overlay of mRNA numbers for each animal that was analyzed, and panel B compares the sliding aver-ages _{± the standard error as a function of time for both strains. Due to the} great similarity observed, we concluded that in the dig-1 strain Notch ligand expression was not affected by 1° fate adoption by different VPCs.

In order to quantitatively compare the expression behaviors of the two closest VPCs to the AC, the relative expression Q_j of ligand j _∈ {apx-1,lag-2_{}, was defined as}

Q_j = Lj,c− Lj,sc Lj,c+ Lj,sc

, (4.1)

were L_j,c is the level of ligand j in the VPC closest to the AC, and L_j,sc is the level of ligand j in the second closest VPC to the AC. Qj varies from -1, when the second closest VPC is the only one expressing mRNAs, to 1, when only the closest VPC shows expression. When its value is zero, both cells contain the same amount of mRNAs. Animals were categorized by dividing the range of Qj in three intervals represented by the colors shown in figure 4.4A. Green includes animals in which Q_j is close to 1, pink when it is close to 0, and purple

(31)

Chapter 4 Experimental results and discussion

A B

C

apx−1 lag-2 Nuclei

AC P5.p P4.p P6.p AC P5.p P4.p P6.p AC P5.p P4.p 10µm

Figure 4.5 A Early L3 dig-1 worm with a correct position of the AC and expression of Notch ligands exclusively in the closest VPC. B Early L3 dig-1 worm with a severely misplaced AC and symmetric expression of Notch ligands. C Late L3 dig-1 animal with a correct cell configuration and expression restricted to the closes VPC. Blue corresponds to cell nuclei stained with DAPI, green to apx-1 mRNAs and red to lag-2 mRNAs. The lines that mark the cell contours are representative and do not correspond to the actual cell membranes.

when it is close to -1.

The obtained relative expression of both ligands for the analyzed animals is shown in figure 4.4B. Each point corresponds to one of the animals con-tained in figure 4.1A and, in addition, the color of the marker indicates the relative AC position for that animal. The green, pink and purple bands in the background denote the three considered categories of relative expression as indicated in Figure4.4A. Figure4.4C shows the fraction of animals belonging to each category for both ligands per time intervals of one hour.

Figure 4.4C indicates that some early L3 animals show expression in the closest VPC to the AC (see Figure4.5A), and in a significant fraction there is symmetric expression (see Figure 4.5B) as well as predominant expression in the second closest VPC, while this tendency decreases in time until all animals end up showing expression exclusively in the closest VPC (see Figure 4.5C). Moreover, this behavior is stronger in apx-1 than in lag-2, which is likely due to the fact that apx-1 is expressed at low number during early induction, while expression of lag-2 only starts at later stages of vulva induction.

Figure4.4Bsuggested that the degree of expression symmetry is correlated with the degree of AC misplacement, i.e. symmetric expression (Q _{≈ 0) and} expression in the more distant VPC (Q _{≈ −1) are mostly seen in animals} where the AC is severely misplaced since markers for animals with Qj . 0 are predominantly green and blue.

(32)

ex-0 0.1 0.2 0.3 0.4 0.5 −1 −0.5 0 0.5 1 3 6 9 12 0 R Q Hours after start of L3 lag-2 0 0.1 0.2 0.3 0.4 0.5 −1 −0.5 0 0.5 1 R Q apx − 1

Figure 4.6 Relative ligand expression Q of apx-1 and lag-2 as a function of the relative AC position R. Each point is colored according to the age of the worm.

pression Qj as a function of the relative AC position R would yield clustered points at regions in which the AC is correctly positioned (low R) and the ex-pression is limited to the closest VPC (Q_{≈ 1). In contrast, at higher values} of R, where the AC is severely misplaced, most points would be located at variable positions around Q_{≈ 0 where expression is symmetric. Indeed, when} plotting the relative ligand expression versus the AC displacement (figure4.6) most of the points lie above the diagonal from the upper left corner to the bottom right corner. Such findings suggest that Notch signaling is responsible for restricting 1° fate to a single cell, even when it is not initially the closest to the AC, and after this fate has been established, cells migrate to finally acquire the correct configuration.

4.2 Inhibition of Ras signaling

The data that we have obtained so far shows that there is a rearrangement of cells during vulva induction, with the AC eventually centering on the VPC with the highest amount of Notch ligands. A possibility is that this correction results from an active mechanism that depends on 1° cell induction, but it might also be simply the result of an independent mechanism such as a passive rearrangement due to the growth of cells within the animal. To address this issue, we proceeded by perturbing the Ras signaling pathway, thus preventing VPC induction. If the rearrangement of cells is dependent on induction, we expect the fraction of animals presenting correct AC positions as well as the variability among animals to remain relatively constant in time. On the other hand, if no active mechanism is involved, we expect to observe similar results as those shown in Figure4.1 for dig-1 animals.

(33)

and consequently, mutations in many of the involved genes result in lethal phenotypes and the animals do not reach the L3 stage [18]. Therefore, we sought for mutant strains that could develop normally and at the same time could prevent vulva induction. We found two strains with mutations in the sos-1 and lin-3 genes that could serve for such purpose.

0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 120 apx − 1 mRNAs in P6.p 0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 120 lag − 2 mR NAs in P 6. p

Hours after start of L3 N2 (WT) 0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 120 0 1 2 3 4 5 6 7 8 9 10 11 12 0 20 40 60 80 100 120

lin-3 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 1 2 3 4 5 6 7 8 9 10 11 12 0 1 2 3 4 5 6 7 8 9 10 11 12

dig-1;sos-1 (25 ºC) apx − 1 mRNAs in P5.p lag − 2 mR NAs in P 5. p apx − 1 mRNAs in P6.p l ag − 2 mR N As in P 6. p 0 0 20 40 60 80 100 120 20 40 60 80 100 120 0 1 2 3 4 5 6 7 8 9 10 11 12 0 1 2 3 4 5 6 7 8 9 10 11 12

dig-1;lin-3 apx − 1 mRNAs in P5.p l ag − 2 mR N As in P 5. p A D C B

Figure 4.7 Number of apx-1 and lag-2 mRNAs expressed in P6.p in N2 and lin-3 mutant (A and C), and in P5.p in dig-1;sos-1 mutant at 25°C and dig-1;lin-3 mutant (B and D).

4.2.1 Inhibition of Ras signaling cascade on VPCs

Our first approach to inhibit the transduction of the inductive signal was to target an important step involved in the Ras pathway. We focused on sos-1, a gene that encodes a protein that is crucial for the LET-23 receptor to re-lay the inductive signal to LET-60, the next component of the pathway [18] (see Figure 1.4B). A temperature sensitive (ts) allele of this gene is available, allowing SOS-1 to function normally at 20°C while impairing its function at

(34)

25°C. Therefore, by allowing the animals to grow at 20°C before L3 and then shifting them to 25°C, they can develop normally and the Ras signaling path-way can be impaired later at the stage of interest. A dig-1;sos-1(ts) double mutant had been previously generated by J.S. van Zon and could be readily used. We found that animals were viable when put at 25°C after hatching; they developed normally except for the lack of a vulva which was expected since the transduction of the inductive signal was inhibited.

We first assessed the extent to which Ras signaling is decreased in this strain by measuring Notch ligand mRNA levels. As presented in figures4.7A

and4.7B, in most of the animals the amount of both apx-1 and lag-2 mRNAs was decreased in dig-1;sos-1 at 25°C compared to N2 animals (whose levels were previously shown to be similar to the dig-1 single mutant). However, a substantial fraction still showed Notch ligand expression, albeit at a lower level. 0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.1 0.2 0.3 0.4 0.5 R

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.2 0.4 0.6 0.8 1 Fr acti on of worms

A

B

Figure 4.8 A Relative AC positions obtained for dig-1;sos-1 worms at 25°C throughout the course of L3. The black line represents a sliding average with a window size of two hours and evaluated every 0.5 hours. The gray area represents the average_{± the standard deviation. B Fraction of} worms that belong on each displacement category per time intervals of one hour.

Despite the few worms that showed close to normal levels, we continued by evaluating cell positions and the results are shown in Figure4.8A. Remarkably, the corresponding fractions of animals per category in Figure4.8Bdo not show

(35)

0 1 2 3 4 5 6 7 8 9 10 11 12 125 150 200 175 Worm length ( μm ) 50 70 90 110 Distance b etween P3.p and P7.p ( μm )

0 1 2 3 4 5 6 7 8 9 10 11 12

dig-1 20ºC dig-1;sos-1 20ºC dig-1;sos-1 25ºC dig-1 25ºC

A B

Figure 4.9 Length of animals (A) and distance between P3.p and P7.p (B) as a function of time (gonad length) for dig-1 and dig-1;sos-1 worms at 20°C and 25°C. Lines represent sliding averages with a window size of two hours and evaluated every 0.5 hours. The shaded areas represent the average_{± the standard error.}

a clear correlation with time as was the case in dig-1 animals.

However, we noted that in dig-1;sos-1 animals, VPCs started undergoing cell divisions at early stages and no animals with undivided cells were observed after 10 hours. In contrast, such cell divisions are observed only at later time-points in wild-type and dig-1 animals. Additionally, we noticed that the AC specification also occurred at earlier stages. We reasoned that dig-1;sos-1 an-imals might not follow the same correlation between gonad length and time that was obtained for wild-type animals (described in section 2.1.1). This difference could be due to the different temperatures at which animals were grown and/or to a different rate of gonad growth brought by the sos-1 muta-tion. Consequently, the gonad length might not provide an equivalent measure of time for both strains, making any comparison between them uncertain.

We then looked for other anatomical markers that could provide us a way to determine the difference in the relationship between gonad length and time in dig-1;sos-1 animals with respect to wild-type, with the aim of obtaining an equivalent measure of time. We focused on two properties that are known to increase in time, namely the length of the animals and the distance in between VPCs.

The two properties were assessed in both dig-1 and dig-1;sos-1 animals, each at 20°C and 25°C. Figure 4.9 shows sliding averages of the worm length (A) and distance between P3.p and P7.p (B) as a function of gonad length. The data shows that the relationship between animal and gonad growth is not altered in the different strains and temperatures considered, suggesting that the two new proposed anatomical markers might not provide an equivalent measure of time either.

(36)

The inconclusive results led us to search for another strain in which Ras signaling is impeded in a tissue specific manner so that other Ras-requiring mechanisms that could be involved in gonad and animal growth would remain unaffected.

4.2.2 Inhibition of lin-3 production in the AC

Among the many strains that have different mutations in the lin-3 gene, there is one in which the only apparent defect is the failure of animals to develop a vulva. This lesion involves a point mutation in a non-coding region that serves to promote lin-3 expression exclusively in the AC [28]. We did not consider this strain before since it has an incomplete penetrance, that is, 11% of the worms do develop a proper vulva and 13% show a ventral protrusion suggesting a low level of induction [29].

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.1 0.2 0.3 0.4 0.5 R

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.2 0.4 0.6 0.8 1 Fr acti on of worms

A

B

Figure 4.10 A Relative AC positions obtained for dig-1;lin-3 animals throughout the course of L3. The black line represents a sliding aver-age with a window size of two hours and evaluated every 0.5 hours. The gray area represents the average _{± the standard deviation. Arrow} indi-cates the animal pictured in4.11. B Fraction of animals that belong on each displacement category per time intervals of one hour.

When we examined Notch ligand expression in the lin-3 strain (Figure

4.7C) we noticed that the levels were remarkably lower than in the dig-1;sos-1 animals. Furthermore, the timing of both, AC specification and VPC division,

(37)

is consistent with what is observed in N2 and dig-1, suggesting that the pre-vious timing issues were caused by impaired Ras signaling in other cells than the VPCs.

Once we knew that ligand levels were sufficiently low, and thus induction occurred infrequently in lin-3 animals, we proceeded to generate the double mutant dig-1;lin-3 to compare the dynamics of cell positions throughout L3 with the dig-1 single mutant. The procedure used to generate this strain is described in appendix B.

As shown in Figure 4.10A, the relative AC position was variable at both early and late stages. Not only did the average remain relatively constant with time at R _{∼ 0.3 (AC mildly displaced), but the standard deviation was} similarly high at all stages. When plotting the fraction of animals per category (Figure 4.10B), the tendency of the cell positions to shift towards a correct configuration is not present, similarly to what we had observed in the dig-1;sos-1 mutant. Figure 4.11 shows a late L3 dig-1;lin-3 worm with a severely misplaced AC, a configuration that we never found on N2 or dig-1 single mutant animals.

AC

P4.p P5.p

10µm Nuclei

Figure 4.11 Late dig-1;lin-3 L3 animal with a severely misplaced AC. Blue corresponds to cell nuclei stained with DAPI. The lines that mark the cell contours are representative and do not correspond to the actual cell membranes.

Finally, a comparison between the sliding average _{± the standard error of} the relative AC position of dig-1 and dig-1;lin-3 animals (Figure 4.12) shows a considerable difference in behavior towards the later stages. This conclu-sively shows the presence of an active migration process dependent on 1° fate induction and Ras signaling.

The question of whether it is the VPCs that migrate towards the AC or the other way around still remained. However, we noted that in the studies performed with lin-12 mutants described in section2.1.2, some animals showed adjacent 1° cells. In these cases, two or three 1° cells were clustered together beneath the AC, suggesting it is the 1° VPCs that move towards the AC.

(38)

0 1 2 3 4 5 6 7 8 9 10 11 12 0 0.1 0.2 0.3 0.4 0.5 dig-1

R

dig-1;lin-3

Figure 4.12 Average relative AC position as a function of time for dig-1 (blue) and dig-1;lin-3 (purple) worms. Lines represent sliding averages with a window size of two hours and evaluated every 0.5 hours. The shaded areas represent the average± the standard error.

4.3 Inhibition of Notch signaling

We now turn to inhibit the Notch pathway to test whether lateral inhibition by Notch signaling is essential for the restriction of 1° fate to a single VPC when the AC is misplaced. Notch signaling plays an important role in the specification of the AC and its inhibition results in the formation of two or more ACs [19]. As mentioned earlier, animals with two ACs usually present either two or three adjacent 1° cells but it is unclear whether this is a result of an increased amount of secreted LIN-3. Therefore, to determine the role that Notch signaling plays in animals with normal amounts of LIN-3, it is important to perturb the pathway in a tissue specific manner or at least in a way that does not interfere with the AC development.

No available mutant is known to completely eliminate Notch signaling in VPCs without side-effects in the AC. There exist temperature sensitive mutations in the Notch receptor gene lin-12 which could serve to allow the animals to develop normally until the AC is specified and only later inhibit the receptor. However, the AC specification process and the start of lateral signaling among VPCs occur both within a short time interval, which would make it difficult to start inhibition at the precise moment.

Therefore, we looked for mutants that weakened Notch signaling specifi-cally in vulva induction and we found two candidates: lip-1 and osm-11. Dur-ing normal induction, the phosphatase LIP-1 is produced in P5.p and P7.p when Notch signaling is activated and its effect is to inactivate the Ras path-way preventing 1° induction (see Figure 1.4B). Loss of function mutations in lip-1 show a wild-type phenotype and only when the activity of the Ras pathway is increased, some animals show adjacent 1° cells [30].

Robustness of vulva induction to anatomical variability in C. elegans

MSc Physics

Physics of Life and Health

Master Thesis

Robustness of vulva induction to anatomical

variability in C. elegans

Abstract

Table of contents

Introduction

1

1.1

Robustness

1.2

C. elegans

}

A

B

A

B

Previous work

2

2.1

Notch ligand expression during vulva

induc-tion

A

B

2.2

Anatomical variability in wild-type animals

A

C

B

2.3

Remaining questions and approach

Techniques and methods

3

3.1

Handling and maintenance of C. elegans

3.2

Single-molecule FISH

}

}

}

Experimental results and

discussion

4

4.1

Increasing anatomical variability

4.2

Inhibition of Ras signaling

A

B

A

B

4.3

Inhibition of Notch signaling