Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 What are the Challenges in Studying the Centrosome and Why Use Drosophila?

Like chromosomes and yeast spindle pole bodies (SPB), centrosome numbers in the cell is strictly controlled. Control of centrosome numbers is achieved by a process of duplication in which the preexisting structure is used as a means to ensure that only a single new structure is formed. Centrosomes consist of two centrioles surrounded by PCM. Centrosome duplication starts after the two preexisting centrioles separate slightly and a new centriole forms near each of the preexisting centrioles. Centrosome duplication is concluded when each centriole pair completely separates, along with some of the PCM of the original centrosome. In each centrosome, the older centriole is also known as the mother centriole and the new centriole is known as the daughter centriole.

The process of centrosome duplication is conceptually similar to how DNA and yeast SPB duplicate (Fig. 1.1). DNA duplication starts by splitting into two strands, which then serve as a template to create a new strand. Yeast SPB duplicate by splitting into two halves. Each half contains a structure known as the half bridge, which then serves to template the formation of another half bridge (Jaspersen and Winey 2004; Jones and Winey 2006). Since, like DNA and SPB duplication, centrosome duplication maintains one preexisting element and creates one new element, it is thought that centrosomes duplicate in a semi–conservative manner.

Fig. 1.1
figure 1

Models for duplication a Yeast spindle pole bodies (SPB Duplication); the half bridge of the preexisting (mother) SPB serves as a template and nucleation site for the new (daughter) SPB that is formed in parallel. b Centriole Duplication; the new (daughter) centriole is formed perpendicularly to the preexisting (mother) centriole and at a significant distance from the surface of the preexisting centriole

Full size image

Although many of the proteins involved in centrosome duplication have been recently identified, the critical question of how the centrioles duplicate remains elusive. Very little is known about the overall mechanism of centriole duplication (Azimzadeh and Marshall 2010; Nigg and Raff 2009). It has become increasingly accepted that centriole duplication does not involve a templating mechanism like in DNA and SPB duplication.

Several lines of observation suggest that de novo formation of the new centrioles takes place at the vicinity of the preexisting centriole.

  1. i)

    The new centriole is formed perpendicularly to the preexisting centriole.

  2. ii)

    The new centriole forms at a distance of about 30 nm from the surface of the preexisting centriole (Anderson and Brenner 1971; Phillips 1967),

  3. iii)

    The new centriole can have a very different structure from the preexisting centriole (Phillips 1967).

  4. iv)

    Under certain conditions, centrioles can form in the absence of a preexisting centriole (Fulton and Dingle 1971; Rodrigues-Martins et al. 2007b)

  5. v)

    Several new centrioles can be induced to form simultaneously around the preexisting centriole by overexpressing centriolar components (Kleylein-Sohn et al. 2007).

  6. vi)

    No proof is available of an intrinsic asymmetry around the preexisting centriole before the onset of centriole duplication

These observations suggest that a yet unidentified mechanism inherent to the centrosome assures that only one new centriole is formed near a preexisting centriole (Sluder and Khodjakov 2010).

Analysis of centriole duplication is challenged by a combination of factors. Centrosomes are essential for development in animals. Centrosomes and centrioles are in low-abundance, found only in one or two copies per cell, thus challenging biochemical approaches. Furthermore, new centriole intermediates are few, small, short-lived, and form too close to the preexisting centriole to be observed as a distinct entity, making it extremely challenging to study centriole intermediates by traditional light microscopy (the internal structure of the centriole is beyond the resolution of standard light microscopy).

Despite these challenges, many proteins involved in centriole duplication have been identified over the past 15 years. The recent identification of many proteins involved in centriole duplication opens new ways to overcome these barriers. One commonly used way is to overexpress the centriolar protein, usually in immortalized cells (in vitro), and observe the consequences using microscopy (Dzhindzhev et al. 2010; Gopalakrishnan et al. 2010; Tang et al. 2009). While sometimes informative, interpreting overexpression data is problematic due to the fact that proteins are studied at non-physiological levels. Also, immortalized cells often have abnormal centrioles, suggesting they already carry mutations that prevent normal centriole duplication. Therefore, to balance these limitations, it is important to develop approaches to study centriole duplication in vivo, when proteins are expressed at physiological levels.

For a number of reasons, Drosophila is ideal for developing such a balanced approach for studying centriole duplication and centrosome biogenesis:

First: mutants defective in centrosome biogenesis are available (see Table 1.1). In flies, even null mutations in essential centrosomal proteins are not embryonic lethal and the fly can often develop to maturity. This is due to maternal contribution which allows the embryo to form centrosomes when they are critical for development, namely during early embryonic development. Later during pupal development when the adult fly is forming, maternal contribution becomes depleted but centrosomes are no longer essential for development. This allows extensive characterization of defective centrosome biogenesis in the testes and sensory neurons of pupae (Basto et al. 2006; Blachon et al. 2009; Blachon et al. 2008; Mottier-Pavie and Megraw 2009).

Table 1.1 Drosophila genes involved in centrosome biogenesis
Full size table

Second: techniques are available to introduce newly-engineered proteins with modified capabilities into a null mutant background, allowing their specific function to be studied with expression at near physiological levels and in the absence of the wild-type protein (Blachon et al. 2008; Gopalakrishnan et al. 2011). This is especially useful in the study of centrosomes, as many centrosomal proteins form multiprotein complexes and may have more than one function. The ability to engineer a mutant that is deficient in one or limited interactions is very insightful in identifying the separate functions mediated by a centrosomal protein (Gopalakrishnan et al. 2011).

Third: it is possible to biochemically isolate centrosomes and centrosomal complexes from Drosophila embryos to study them ex vivo (Gopalakrishnan et al. 2010; Gopalakrishnan et al. 2011; Kellogg and Alberts 1992; Moritz et al. 1995). This allows one to study protein interactions under near physiological conditions. This also opens a window to use purified centrosomal proteins, structures, and complexes in cell-free experiments that can investigate the individual steps in centrosome duplication. Ultimately, this can theoretically allow centrosome duplication and function to be reconstituted using purified components.

Finally: Drosophila centrosomes are formed using conserved proteins and the overall structure of Drosophila centrosomes is very similar to that of other organisms, suggesting that the basic mechanisms of centrosome duplication used in Drosophila are similar to those used in other organisms.

2 Centrosomes in Drosophila Development

Centrosomes in Drosophila were studied in some detail in a context of several developmental processes. In this chapter, we will focus on four processes. The first two processes take place during early embryonic development: (1) Fertilization, and (2) Syncytial blastoderm formation. The next two processes occur in differentiated cell types and can be studied during pupal development: (3) Sensory neuron differentiation, and (4) Spermatogenesis. We will summarize key features of the centrosome in each of these developmental processes, highlighting unique properties that have provided insight into the biology of the centrosome.

2.1 Fertilization

Fertilization is the process by which the sperm (male gamete) and oocyte (female gamete) are fused to form a zygote, the first cell of a new organism. In general, a key step in fertilization is the migration of sperm and oocyte pronuclei toward each other and their subsequent fusion (Fig. 1.2).

Fig. 1.2
figure 2

Fertilization during mitosis of the zygote (a), the giant centriole (b) and a second smaller centriolar structure (c) are observed at the two poles. Note that the two pronuclei (blue) are not mixed and they divide separately in parallel. A low magnification image with small inset squares outlines the approximate positions of the centriolar structures, which are shown under a higher magnification in (b and c). Embryos were stained with rat anti-α-tubulin (red), and anti–N-ter-Asl (purple); 4’-6-diamidino-2-phenylindole (DAPI) stains DNA

Full size image

It is generally recognized that in most animals, including Drosophila, the oocyte does not contain centrioles (Krioutchkova and Onishchenko 1999; Manandhar et al. 2005; Sun and Schatten 2007). Instead, oocytes have acentriolar centrosomes or microtubule organization centers that participate in female meiosis and in the formation of the female pronucleus (Megraw and Kaufman 2000). While the oocyte does not appear to have centrioles, it does contain a large amount of centriolar and PCM proteins within its cytoplasm, enough to form 213 centrosomes (Rodrigues-Martins et al. 2007b). These proteins, contributed by the mother via the oocyte (maternal contribution), are sufficient to support centrosome duplication in early embryonic development, a time when the embryonic genome is not yet fully involved in producing the proteins necessary for development.

On the other hand, the Drosophila sperm contains two centriolar structures. The first, termed the “giant centriole” (due to its exceptional length) resembles the distal centriole found in vertebrate sperm and functions to nucleate the sperm flagellum (Friedlander and Wahrman 1966; Fuller 1993; Krioutchkova and Onishchenko 1999; Manandhar et al. 2005; Sun and Schatten 2007). A second centriolar structure associates with the giant centriole and is termed the proximal centriole-like (PCL) structure due to the fact that, like the vertebrate sperm proximal centriole, it does not form a flagellum (Blachon et al. 2009) (Fig. 1.3).

Fig. 1.3
figure 3

Spermatid with a giant centriole (GC) and a PCL. Left, giant centriole and PCL labeled by Ana-1-GFP (purple). Right, cartoon depicting the relative location of the sperm centriolar structures relative to the sperm head (H) and tail (T). (The panel on the Left is modified from Fig. 1.2b in (Blachon et al. 2009)

Full size image

Upon fusion of the sperm and oocyte, the sperm giant centriole recruits PCM proteins from the surrounding cytoplasm and forms a centrosome. The zygote centrosome acts as a microtubule organization center and assembles an aster—a star-like structure consisting of microtubules. These asters are thought to play a role in bringing together the female and male pronuclei and in orchestrating the first cell division (Callaini and Riparbelli 1996). In support for the critical role of centrosomes in zygote biology, it has been reported that interfering with centrosome biogenesis after fertilization inhibits zygote development (Dix and Raff 2007; Stevens et al. 2007; Varmark et al. 2007).

The role of the PCL after fertilization is currently not known; however, one attractive hypothesis is that the PCL later becomes the second centrosome of the zygote. The observation that two centrosomes can be observed after fertilization in mutant oocytes under conditions that block centrosome duplication supports this hypothesis (Stevens et al. 2007).

2.2 Syncytial Development

Drosophila early embryonic development takes place in a syncytial blastoderm, a large cell containing many nuclei. In this developmental stage, the embryo undergoes 13 rounds of nuclear duplication and division without forming individual cells surrounded by a plasma membrane; each of these rounds is called a “nuclear cycle”. For simplicity, we will refer to each dividing unit that includes a nucleus and its associated centrosomes as a “cell”. At the end of syncytial blastoderm development, the nuclei are partitioned into separate cells where they are surrounded by a plasma membrane (cellular blastoderm stage). This partitioning, termed cellularization, is mediated by actin, which forms a cleavage furrow (the structure that mediates the separation of daughter cells).

The syncytial blastoderm is an excellent system where one can do live imaging of centriole duplication in real time (Fig. 1.4). In the embryonic syncytium, the nuclear cycle is very rapid (~10 min) and many centrioles duplicate synchronously. The nuclear cycle is comprised of two phases: synthesis and mitosis. Early in synthesis phase, each cell has two centrosomes, each containing one centriole surrounded by PCM (Callaini and Marchini 1989; Callaini and Riparbelli 1990, 1996; Riparbelli et al. 1997). During synthesis phase, the centriole within each of the two centrosomes duplicates, generating a new centriole at the vicinity of the older centriole. At the onset of mitosis, the centrosomes move to opposite poles such that each future daughter “cell” will inherit one of the of the mother “cell” centrosomes. During late mitosis, each of the centrosomes that are associated with one of the nuclei splits into two centrosomes. As a result, each of the daughter “cells” inherently contains two centrosomes, each with one of the centrioles from the original centrosome (Fig. 1.5).

Fig. 1.4
figure 4

Live imaging of centriole duplication in the early development of an embryo expressing the centriolar marker Sas-6-GFP. The start of centriole separation marks the splitting of the centrosome. The left panel shows low magnification of the centrioles, organized in the right panels by time. The white square in the panel on the right corresponds to the centrioles displayed in the left panels. A line distinguishes each centriole. S. B produced this picture

Full size image
Fig. 1.5
figure 5

Centriole duplication and centrosome separation in the syncytial blastoderm. The embryo expressed the centriolar marker Sas-6-GFP (green) and was stained with an anti-Asl antibody; Asl is a component of the PCM that is found near the centriole (red). a In early S phase, each of the two centrosomes contain the PCM protein Asl and have a centriole labeled by Sas-6-GFP. b During S phase each of the centrioles duplicates to form a daughter centriole (Dc). The daughter centriole is marked with Sas6-GFP but not Asl. c–f Only one of the centriole pairs is shown. The mother and daughter centrioles start to separate and the daughter centriole accumulates Asl, finally leading to the formation of two centrosomes. Note that anti-Asl antibody also lightly labeled the nucleus. S. B produced this picture

Full size image

The centriole in the Drosophila syncytial blastoderm has a very intriguing structure. Unlike classic centrioles, which are made of nine triplets of microtubules, centrioles in the syncytial blastoderm are made of nine doublets of microtubules. In addition, syncytial blastoderm centrioles are shorter than classic centrioles: ~200 nm long instead of 400–500 nm long observed in vertebrates. Finally, the centrioles of the syncytial blastoderm have a structure known as the “cartwheel” within their core, which in vertebrate cells is characteristic of a young, developing centriole (procentriole) and is absent from mature centrioles. This raises the hypothesis that syncytial blastoderm centrioles are centriolar structures arrested in the procentriole stage. It is possible that since the syncytial blastoderm nuclei divide rapidly and there is no need for cilia formation (see below), the centrioles do not have the time, nor the need, to develop into their mature states (Fig. 1.6).

Fig. 1.6
figure 6

The fly early embryonic centriole resembles a procentriole. (The fig is modified from Fig. 1 in (Avidor-Reiss 2010)

Full size image

2.3 The Drosophila Zygote and Syncytial Blastoderm Develop Using Proteins Generated in the Mother

Protein deposition in the oocyte that supports early embryonic development is called maternal contribution. An important implication of the presence of maternal contribution in Drosophila development is that, despite the fact that an embryo may be genetically homozygous for a mutation in an essential centrosomal gene, it will still contain the wild-type protein, allowing it to produce normal centrosomes as long as the maternal contribution persists. As a result, studying those mutants cannot reveal the role of centrosomes in early embryogenesis. Indeed, studies using homozygous mutants for an essential centrosomal gene, have demonstrated that the fly embryo can develop normally (Basto et al. 2006; Blachon et al. 2008; Rodrigues-Martins et al. 2007a).

Investigating the role of centrosomes in early embryogenesis requires the study of an embryo that is produced from an oocyte generated in an environment that is also mutated. Since flies with mutations in essential centrosomal proteins are unable to walk, mate, or lay eggs (see below), it is not possible to use embryos produced by homozygous females. However, this obstacle can be overcome using other approaches.

One way is to study centrosomal proteins that are essential for aspects of centrosome function but are not necessary to produce a fertile female. For example, mutations in centrosomin (Cnn) result in flies that are viable but female sterile (Megraw et al. 1999; Vaizel-Ohayon and Schejter 1999). Cnn is a PCM protein that plays an important role in PCM formation and is required for the centrosome’s activity as a microtubule organization center (Li and Kaufman 1996). Studies of cnn mutants in early embryogenesis reveal an impairment of several aspects of embryo development that depend on the function of the cytoskeleton (Megraw et al. 1999; Vaizel-Ohayon and Schejter 1999). In particular, it appears that Cnn is essential for the organization of actin into cleavage furrows. New information suggests that the centrosome functions as a site where Cnn interacts with Centrocortin, a protein that is required for cleavage furrow formation and is localized both to centrosomes and to cleavage furrows (Kao and Megraw 2009). It is therefore possible that the centrosome functions as a signaling hub within the cell. At this signaling hub, proteins can interact in order to integrate information and later move to other domains in the cell where they execute their function (Alieva and Uzbekov 2008; Wang et al. 2009).

Another way by which to study centrosomes in early embryogenesis is to study hypomorphic mutations of centrosomal proteins that are essential for centrosome formation and produce a female that can mate. In this case, the studied protein is partially functional and missing an activity that is essential specifically for centrosome function during embryogenesis. An example of one such mutation is asl 1, in which the C-terminus of the essential centrosomal protein Asterless (Asl) is truncated (Blachon et al. 2008). While flies homozygous for the severe loss-of-function allele asl mecD are unable to walk and mate, asl 1 generates viable females that can lay eggs. Analysis of embryonic centrosomes generated from an asl 1 female finds that they initially form asters, but these asters are not stable and later fall apart ((Varmark et al. 2007) and S. B. unpublished data). In addition, pronuclei fusion does not take place and embryo development is arrested at the zygote stage. Similar results were obtained when a hypomorphic mutation of the Drosophila spd-2 gene was studied, while severe loss-of-function spd-2 mutants are unable to walk and mate (Dix and Raff 2007; Giansanti et al. 2008).

One can also study centrosomal proteins that are essential for centrosome formation and produce a female that can mate using genetic tricks. One way to do so is to make germline clones that lack any particular essential centrosomal protein from maternal contribution (Stevens et al. 2007). This is done using the dominant female sterile (DFS) technique (Chou and Perrimon 1996), a variant of FLP-FRT recombination. In this case, recombination is induced in larvae via a heat shock-inducible flippase (FLP); as adults, the fly produces homozygous mutant oocytes that lack the essential centrosomal protein. Study of oocytes that are mutant for sas-4 or sas-6 after they are fertilized with a wild-type male finds that they possess two centriolar structures (presumably the giant centriole and PCL) (Stevens et al. 2007). These centriolar structures can nucleate centrosomes and can undergo few nuclear cycles before embryogenesis is arrested presumably because the centrosomes cannot duplicate. This suggests that the PCL can form an independent centrosome and demonstrates that centrosomes are essential for syncytial blastoderm development (after their role in zygote pronuclei fusion).

An alternative method to study centrosomes in the syncytial blastoderm is to inject it with an antibody for a particular centrosomal protein and observe the consequence (Conduit et al. 2010). A potential problem with this approach is that when a centrosomal protein forms a complex, binding of an antibody may not only inactivate its intended protein target, but may also inactivate other proteins found in the complex.

3 Centrosomes in Differentiated Cells

3.1 Sensory neuron differentiation

In Drosophila, the first cells that develop cilia are the type I sensory neurons (Fig. 1.7). These neurons mediate the reception of mechano- and chemo-sensory information and are found in both larvae and the adult fly. One subtype of these sensory neurons is found on the cuticle of the adult fly and mediate touch sensation (Fig. 1.7) (Keil 1997). These neurons are bipolar sensory neurons that extend a dendrite with sensory cilia at their tip in one direction. The sensory cilia are attached to a cutaneous structure termed the bristle. When the bristle moves due to mechanical stimuli, the mechanosensory transduction machinery found in the cilia is activated and a mechanoreceptor current is generated (Fig. 1.8). This current produces an action potential that is delivered to the brain, transmitting information regarding touch sensation or proprioception (Avidor-Reiss et al. 2004; Kernan et al. 1994; Walker et al. 2000).

Fig. 1.7
figure 7

Drosophila mechanosensory neuron morphology. a The neuron cell body is filled with tubulin-GFP. The dendrite is lightly labeled by tubulin-GFP. At the dendrite tip, the cilium is strongly labeled by tubulin-GFP. Red labels cuticle structures, including the bristle. Blue labels the Transition zone vicinity. b Diagram of sensory cell dendrite and cilium. (The a panel is modified from Fig. 7c in (Avidor-Reiss et al. 2004)

Full size image
Fig. 1.8
figure 8

Drosophila mutants with centriolar or cilia defects are mechanosensory defective. Displacing a bristle 30 um for one second (a) generates a mechanoreceptor current in control flies that adapts over the course of the stimulus (b). In contrast, mutations that affect centriole formation (c and d) and thus cannot form mechanosensory cilia have no mechanoreceptor current. (The a and b panels are modified from Fig. 1.4c in (Avidor-Reiss et al. 2004)

Full size image

The sensory neuron is a product of asymmetrical cell division and it inherits a centrosome with two centrioles (Gomes et al. 2009; Keil 1997; Seidl 1991). After the sensory neuron generates a long dendrite, the two centrioles are reorganized and are found in tandem, one of which is attached to a vesicle. This reorganized structure appears to migrate along the dendrite until it reaches the distal end, where the associated vesicle fuses with the plasma membrane to form the sensory cilium (Seidl 1991). The sensory cilium is composed of a transition zone, also called the connecting cilium, and the sensory cilia proper, also called the outer segment (Fig. 1.7b).

3.2 Spermatogenesis

Spermatogenesis is the process that takes place in the testes to form mature male gametes and begins when a sperm stem cell divides asymmetrically to form another stem cell and a progenitor spermatogonium. The spermatogonium divides 4 times to form 16 spermatocytes. These spermatocytes grow to ~30 times their original size and ultimately undergo two cycles of meiosis to generate 64 spermatids. The spermatids, which are first round, undergo a dramatic differentiation program, called spermiogenesis. The completion of this differentiation program results in the formation of a sperm cell that is ~2 mm long, a length comparable to that of the fly itself.

Centrosomes in the Drosophila testes have several interesting properties:

First, unlike the syncytial blastoderm, the centrosome and centrioles of the adult testes are similar to their vertebrate counterparts (Tates 1971; Tokuyasu 1975). These centrioles have nine triplet microtubules (Fig. 1.9k). This normal centriolar structure correlates with the ability of the spermatogenic centrioles to form cilia and suggests that in ciliated cells, the centriole needs to develop into its mature state. In this regard, sensory neurons that have cilia also contain centrioles with triplet microtubules (Keil 1997).

Fig. 1.9
figure 9

The giant centriole (proximal centriole) and primary cilium (distal centriole) of fly spermatocytes: A-J serial section electron microscopy analysis of a pair of giant centrioles organized in an orthogonal relationship. k The cross-section of the daughter centriole from c is magnified to demonstrate triplet microtubules. j The last cross-section is highlighted to depict the presence of irregular numbers of doublet microtubules in the primary cilium. TF, transitional fibers connecting the centriole to the plasma membrane; TM, triplet microtubules; DM, doublet microtubules; CM, cilium membrane; pc and dc according to Tates: pc, proximal centriole or basal body and dc, distal centriole

Full size image

Second, in spermatogenesis, the centrosomes form two types of cilia. During spermatocyte growth, each of the spermatocyte’s four centrioles forms a primary cilium-like structure of unknown function (Fig. 1.9b) (Tates 1971). Later in spermiogenesis, each of these primary cilia is modified to form a motile cilium—the sperm flagellum.

Third, male meiosis is absolutely dependent on centrosomes. Flies that do not have functional centrosomes fail to accurately separate genetic material and mitochondria (Fig. 1.10), one of the reasons why centrosomal mutants are male sterile. It is currently unclear why centrosomal defects cause abnormalities in male meiosis but do not disrupt mitosis. However, it is possible that male-specific meiotic defects are due to the fact that the centrosomes are associated with a ciliary-like structure that requires centrosomal components for their formation.

Fig. 1.10
figure 10

Round spermatids formed imminently after meiosis, contain a white nucleus (N) and dark mitochondria (M) of similar size (a). Centriolar mutants in the spermatid state form nuclei and mitochondria of variable size (b and c)

Full size image

Fourth, during spermatocyte growth and spermiogenesis, the centriole elongates to ~2.5 um, much larger than centrioles found in any other Drosophila tissue or those of other organisms (Blachon et al. 2009; Mottier-Pavie and Megraw 2009; Tates 1971). At particular stage of spermiogenesis these giant centrioles are surrounded by a long and thick PCM (also referred to as the “centriolar adjunct”). These centrioles provide a very convenient model to study centriole elongation and several mutants that have shortened giant centrosomes have been described. Proteins essential for giant centriole elongation include: Bld10, Poc1, and Sas-4 (Blachon et al. 2009; Mottier-Pavie and Megraw 2009).

Fifth, during spermiogenesis, the spermatid cell forms a centriole precursor-like structure called the PCL (Fig. 1.11a). The PCL has been proposed to be a centriole intermediate that arrests at the stage before the centriolar microtubules are assembled. Therefore, by studying how the PCL forms, one can study the early events in centriole biogenesis (Blachon et al. 2009). PCL formation depends on the function of the centrosomal proteins Plk4, Sas-6, and Asl and it contains the following centrosomal proteins: Sas-6, Ana2, Ana1, Sas-4, Bld10, Cnn, and Asl (Blachon et al. 2009; Mottier-Pavie and Megraw 2009; Stevens et al. 2010b) and (Fig. 1.11b).

Fig. 1.11
figure 11

Asl is essential for PCL formation. a To determine the relationship between the PCL, the giant centriole (c), and the PCM, the localization of Ana1-GFP relative to the centriolar adjunct protein γ-tubulin (Wilson et al. 1997) was analyzed. In early spermatids, γ-tubulin labels the vicinity of the giant centriole along most of its length (i). During the time the Ana1 labelled PCL appears, γ-tubulin assembles a half ring structure around the giant centriole, which touches the PCL (ii). At a later stage, the PCL migrates distally and γ-tubulin labels the PCL. b To determine if Asl plays a role in PCL formation we followed maternally contributed giant centrioles in spermatids of the asl mecD mutant. No PCL is observed near the giant centriole, demonstrating an essential role for Asl in PCL formation. However, while in early asl mecD spermatids γ-tubulin localization appears to be normal, many abnormal γ-tubulin rings are found at the vicinity of the maternally contributed giant centrioles in intermediate asl mecD spermatids. (The a panel is modified from Fig. 1.2c in (Blachon et al. 2009)

Full size image

3.3 The Drosophila Adult Sensory Neuron and Testes Develop Using Proteins Generated During Metamorphosis

Unlike early embryogenesis that utilizes maternally contributed proteins, the adult fly develops during metamorphosis by utilizing proteins synthesized from the genome of the fly itself. Therefore, by studying sensory neurons and testes in the pupa or adult, one can study the full impact of a mutation in a centrosomal protein. Interestingly, flies that have mutations in essential centrosomal proteins and have no centrosomes can develop to adulthood but die soon after leaving the pupa because they cannot stand on their legs, walk, or fly [defects collectively referred to as uncoordination (Kernan et al. 1994)]. This uncoordination results from the fact that flies with centrosomal defects have no mechanosensory cilia and cannot sense the environment or their body parts.

The germline stem cells found in the testes originate from the first group of cells that are generated early in embryonic development (pole cells) (Okada 1998). When germline stem cells divide to form a new stem cell and a spermatogonium, the older centriole (the maternally contributed centriole) stays in the stem cells while the newer centriole is inherited by the spermatogonium (Blachon et al. 2008; Yamashita et al. 2007). A fly that is homozygous mutant for an essential centriole component and cannot form new centrioles will have functional centrioles in the germline stem cells that are made using maternal contribution, but will lack these centrioles in the later progenitors.

Since germline stem cells in the developing Drosophila testes have two maternally contributed centrioles, some of the first spermatogonium to form can each inherit one maternally contributed centriole. These centrioles then duplicate and elongate during spermatogenesis and end up in the first spermatids to form. By that time maternal contribution of wild-type proteins becomes depleted. Following maternally contributed centrioles of the spermatogonium and later in spermatogenesis allows one to dissect the role of a particular protein under circumstances where a centriole is present (Fig. 1.12). Using this approach, it was found that maternally contributed centrioles require Sas-4 to elongate but not Asl (Blachon et al. 2009; Blachon et al. 2008). On the other hand, formation of the PCL can form in Sas-4 mutants but not in Asl mutants (Blachon et al. 2009) and Fig. (1.11b).

Fig. 1.12
figure 12

Asl is essential for centriole duplication in vivo. Wild-type cells (control) contain both a mother centriole (M) and its daughter centriole (D). Centrioles are labeled by Ana1-GFP (green); PCM assembled around these centrioles are labeled for γ-tubulin (red). In asl loss-of-function mutant cells, the maternally contributed centrioles elongate but its duplication is blocked. Modified from (Blachon et al. 2008). (The fig is modified from Fig. 1.5d in (Blachon et al. 2008)

Full size image

3.4 Phenotypic Characterization of Centrosomal Mutants

The spermatid flagellum is formed in a unique way from that of sensory cilia (Avidor-Reiss et al. 2004; Han et al. 2003). Sensory cilia are formed within a distinct compartment separated from the rest of the cell by a transition zone that is thought to function as the ciliary gate (Betleja and Cole 2010; Craige et al. 2010; Omran 2010; Williams et al. 2011). As a result, sensory cilium formation requires a complex machinery known as the intra-flagellar transport (IFT) (Rosenbaum 2002; Scholey and Anderson 2006). On the other hand, the spermatid flagellum is formed in the cytoplasm and does not depend on IFT, a process called cytoplasmic or non-compartmentalized ciliogenesis (Avidor-Reiss et al. 2004; Han et al. 2003). In flies, this distinction allows phenotypic analysis to rapidly determine if a protein is an essential centriolar protein, if it is important for compartmentalized ciliogenesis, or if it is important for the conversion of a centriole into a basal body.

Mutants in essential centriolar proteins will have the following phenotype that reflects the role of centrioles in their development. Meiosis will be abnormal in males due to the importance of centrosomes in meiosis, sperm will be nonmotile as a result of defects in non-compartmentalized ciliogenesis, and adult flies will be uncoordinated due to defects in compartmentalized ciliogenesis. Such mutants include Sas-6, Sas-4, Asl, Ana-1, Spd-2, and Plk4 (Basto et al. 2006; Bettencourt-Dias et al. 2005; Blachon et al. 2009; Blachon et al. 2008; Giansanti et al. 2008; Martinez-Campos et al. 2004; Rodrigues-Martins et al. 2007a).

Mutations in proteins that are essential for compartmentalized ciliogenesis will not affect meiosis or sperm motility, but will still result in adult uncoordination due to defects in compartmentalized ciliogenesis. Examples include mutation in IFT proteins such as Oseg 1 and 2 as well as IFT88 (Avidor-Reiss et al. 2004; Han et al. 2003).

A third group of centriolar proteins seems to affect compartmentalized and non-compartmentalized ciliogenesis, but do not have an impact on meiosis. This group seems to function in the conversion of a centriole into a basal body. Such mutants include unc (Baker et al. 2004).

4 The Role of the Centrosome

Immediately after the discovery of the centrosome by Flemming (1875) and van Beneden (1876), two major hypotheses regarding its function were postulated. Boveri hypothesized that the centrosome is a cellular organelle found close to the nucleus and with paramount importance in cell division. On the other hand, the Henneguy-Lenhossek theory (1898) claimed that the centrosome and basal body are the same organelle located in two distinct sites, with the centrosome located at the cell center near the nucleus, and the basal body existing at the base of the cilia at the plasma membrane. This theory was the first to emphasize the importance of the centrosome in cilia formation.

These two hypotheses were based only on observations of the centrosome and have remained untested for many years. Recent studies in Drosophila have allowed these theories to be tested directly. A genetic approach that has been employed is elimination of the centrosome using mutants and the study of its effect on cilia formation and cell division. Ideally, one would employ a null mutation in a centrosome-specific component that is absolutely essential for centrosome formation. Studying the phenotype of these mutants can test if the centrosome is essential for cell division and/or cilia formation.

Study of fly mutants that block centrosome formation (sas-4 and asl mecD ) has suggested that flies lacking centrosomes die due to mechanosensory defects caused by a failure to form cilia (Basto et al. 2006; Blachon et al. 2008). These flies also exhibit a failure in male meiosis and their sperm form without an axoneme. However, it is important to note that during larval and pupal development, the centrosome does not play an essential role in cell division; division, though delayed, still takes place. This argues that, at least in larvae, pupae, and adult files, centrosomes and centrioles behave in a way consistent with the Henneguy-Lenhossek theory.

In flies, the oocyte carries a large amount of maternally derived proteins that support the early development of the embryo. Therefore, analyzing the role of the centrosome in the early embryo requires the embryo’s mother or oocyte to be mutant (see above). Interestingly, such studies suggest that centrosomes are vital in the zygote and early embryo. However, whether this is because centrosomes are essential to mediate nuclear division is not clear. Regardless, in this developmental stage, fly cells do not have cilia and it is thus possible that the early embryonic syncytium requires the centrosome for “cell” division in a way that is consistent with the Boveri hypothesize. However, an important caveat in this idea is that centrosomes may have an important function in the early embryo that is neither related to cilia formation nor cell division. For example, it appears that centrosome function is essential for the migration and fusion of the gamete pronuclei and in acting as a signaling hub (see above).

5 Two Pathways of Centriole Formation

A centriole forms by one of two pathways (Anderson and Brenner 1971; Delattre and Gonczy 2004; Loncarek and Khodjakov 2009; Loncarek et al. 2007). In the “acentriolar pathway”, a daughter centriole forms de novo without a preexisting centriole. This occurs when multiple centrioles are required in a cell or under the unusual situation when preexisting centrioles are absent. The acentriolar pathway produces an imprecise number of centrioles. In flies, centrioles can form de novo when certain centrosomal proteins are overexpressed in oocytes or fertilized embryos (e.g., PLK4, Asl, and Ana-2) (Peel et al. 2007; Rodrigues-Martins et al. 2007b; Stevens et al. 2010a). To what extent these centrioles are similar to centrioles that normally form is not clear, but in some cases, they appear to have normal centriolar structures, have the capacity to recruit PCM proteins, and can form astral microtubules. Whether, these induced centrosomes can form cilia or mediate meiosis is unknown. However, overexpression of centrosomal proteins can also form centriolar structures that are clearly abnormal (Gopalakrishnan et al. 2010; Rodrigues-Martins et al. 2007a; Stevens et al. 2010a). The capacity to induce centriole formation by overexpression seems to be tissue-specific and is more prevalent in the oocyte/embryo than in the testes (Peel et al. 2007; Stevens et al. 2009).

Whether centrioles can form de novo in Drosophila melanogaster without overexpression of centriolar components is not clear. It was shown that in other Drosophila species, oocytes assemble a number of cytoplasmic asters after activation with centrioles and centrosomal proteins (Ferree et al. 2006; Riparbelli and Callaini 2003). In addition, in sas-4 null mutants that cannot form centrioles, foci containing centriolar markers appear transiently, suggesting that nascent procentrioles form de novo but fail to develop further in the absence of Sas-4 (Gopalakrishnan et al. 2011).

In the “centriolar pathway”, a preexisting centriole acts as a template to ensure that only a single daughter centriole is formed per cell cycle. However, it does not appear to impart structural information to the daughter (Phillips 1967; Rodrigues-Martins et al. 2007b). Before cell division, centriole duplication under the “centriolar pathway” results in four centrioles; each mother/daughter centriole pair forms a centrosome that migrates to one of the cell’s two poles where it helps orient the mitotic spindle, a structure essential for cell division. Having precisely one centriole pair in interphase and two centriole pairs during mitosis is believed to be critical for proper cell division and having too many centrioles can initiate tumorigenesis (Basto et al. 2008; Cunha-Ferreira et al. 2009; Fukasawa 2007; Ganem et al. 2009).

How the mother centriole ensures that only a single daughter centriole forms is not clear. However, it is readily observed that there is already only one daughter centriole at the vicinity of the mother centriole by the time a procentriole is present. Therefore, an approach to address this question is to study proteins that are involved early in centriole formation when the formation of the procentriole is initiated. In recent years, extensive proteomic, genetic, and bioinformatic studies have identified many of the key proteins critical for centriole formation (Andersen et al. 2003; Avidor-Reiss et al. 2004; Fritz-Laylin and Cande 2010; Gonczy et al. 2000; Keller et al. 2005; Kilburn et al. 2007; Li et al. 2004; Mahoney et al. 2006). Some of these proteins were analyzed in flies and were shown to function early in procentriole initiation (Table 1.1). Some of the important discoveries regarding these proteins that were made using flies are summarized below.

6 Identification of Centrosomal Proteins and Mutants in Drosophila melanogaster

Identification of centrosomal components that play a specific role in centrosome have been hampered by initial difficulties in obtaining sufficient quantities of material for biochemical isolation and a lack of clear phenotypes expected from mutation of essential centrosomal proteins. Over time, several approaches have resulted in the identification of Drosophila proteins important for centrosome biogenesis and function (Table 1.1). This includes the use of anti-centrosomal antibodies, the use of reverse genetic approaches, and the use of forward genetic approaches such as the cloning of genes from sterile, mechanosensory, or PCL mutants.

6.1 Anti-Centrosomal Antibodies

Analysis of the centrosomal gene CP190 (Whitfield et al. 1988) was facilitated by the use of an antibody raised against isolated centrosomes and that were later found to bind CP190. CP60 was identified as a protein that interacts with CP190 (Kellogg and Alberts 1992). CP190 and CP60 localized to the centrosome in a cell cycle-dependent manner and their amount at the centrosome was shown to be maximal during mitosis and was barely detected during interphase. However, CP190 and CP60 are not centrosome-specific and are also found in the nucleus during interphase. A mutant for the CP190 gene was identified using standard genetic approaches and it was found that CP190 is not essential for centrosome biogenesis or function, but its function in the nucleus is essential for fly viability (Butcher et al. 2004). RNAi study of CP60 suggests it is also not essential for centrosome biogenesis or function (Butcher et al. 2004). Orthologs of CP190 and CP60 have not been identified in vertebrates.

6.2 Reverse Genetics

Another way to obtain mutants in centrosomal proteins is to study the Drosophila ortholog of known centriolar proteins in other organisms. This approach benefits from the availability of large collections of mutant flies in identified genes. For example, Sas-4 (Kirkham et al. 2003), Sas-6 (Dammermann et al. 2004; Leidel et al. 2005), Spd-2 (Kemp et al. 2004; Pelletier et al. 2004), Bld10 (Matsuura et al. 2004) and PLK4 (Habedanck et al. 2005) are all conserved centriolar proteins first identified in other model organisms.

6.3 Bioinformatic subtractive screen for ciliary and centrosomal proteins

Several centrosomal proteins were identified in a bioinformatic subtractive screen for ciliary and centrosomal proteins (Avidor-Reiss et al. 2004; Li et al. 2004). Such a screen is based on the idea that only organisms that have cilia should have ciliary genes and is used to identify genes that are conserved in organisms with cilia but are missing in organisms that lack cilia. Since centrioles are only observed in ciliated organisms, this screen can also be used to identify centriolar proteins (Carvalho-Santos et al.; Hodges et al.). Examples of Drosophila centrosomal genes identified using such approaches are poc1 and poc18 (Keller et al. 2005).

6.4 Male Sterile Mutants

The centrosomal mutant asterless (asl) was identified in a study of male sterile mutants (Bonaccorsi et al. 1998). Unfortunately, for nearly 10 years, it was incorrectly believed that Asl was solely important for centrosomal function and aster formation (hence its name: Asterless) while its essential role in centriole formation remained unknown (Bonaccorsi et al. 1998; Giansanti et al. 2001; Oliferenko and Balasubramanian, 2002; Varmark et al. 2007; Wakefield et al. 2001). Nonetheless, it was later demonstrated that Asl is instead essential for centriole duplication and that the absence of aster formation in asl mutants is mainly due to a lack of centrioles (Blachon et al. 2008). This finding came from analysis of asl mecD, a new mutant that was discovered in a mechanosensory mutant screen (see below). Asl is a conserved centrosomal protein known as Cep152 in vertebrates (Blachon et al. 2008; Varmark et al. 2007).

Male sterility is a common phenotype in centrosomal mutants and is a valuable method to identify centriolar mutants. Male sterility in centriolar mutants can arise by several distinct mechanisms:

First, centrosomes are essential for male meiosis and flies with abnormal meiosis will fail to form functional sperm. This type of defect is most commonly diagnosed by examining spermatids immediately after meiosis by light microscopy and observing abnormal numbers and shapes of nuclei and mitochondria (Fig. 1.11) (Bonaccorsi et al. 1998; Li et al. 1998).

Second, the centriole acts in templating the sperm flagellum. As a result, abnormalities or absences of the centriole translate to abnormalities or absences of the sperm flagellum. This type of defect is diagnosed by observing that the fly sperm is not motile, or by electron microscopy analysis where fly sperm cross-section shows a missing or abnormal axoneme (Baker et al. 2004; Blachon et al. 2008; Mottier-Pavie and Megraw, 2009; Rodrigues-Martins et al. 2007a).

Third, centrosomes are essential for mechanosensory cilia formation (Blachon et al. 2009; Blachon et al. 2008; Martinez-Campos et al. 2004). In the absence of normal centrioles, flies have abnormal proprioception and cannot court the female and mate with her.

Fourth, it is possible that an additional mechanism of male sterility is the failure to form a normal PCL. If the PCL becomes one of the zygotic centrosomes, it is expected that a nonfunctional PCL will cause male sterility even if the sperm is delivered to the oocyte and fertilization takes place. In this case, genes whose mutations cause PCL failure are expected to fall into a class of interesting mutants referred to as paternal effect genes (Fitch and Wakimoto 1998).

In some mutants of centrosomal proteins, such as bld10, sterility is caused by an abnormal sperm flagellum (Mottier-Pavie and Megraw, 2009). In other mutants, such as cnn, sterility is a result of meiosis defects followed by abnormal differentiation of the motile sperm (Li et al. 1998). However, in the most severe centrosomal mutants (for example, asl, sas-4, sas-6, plk4, ana1) sterility is caused by a combination of all of these mechanisms; the flies are uncoordinated, fail in meiosis, and do not form flagella (Basto et al. 2006; Bettencourt-Dias et al. 2005; Blachon et al. 2009; Blachon et al. 2008; Martinez-Campos et al. 2004; Rodrigues-Martins et al. 2007a). While, defects in sperm motility, meiosis, or proprioception are not restricted to centrosomal mutations, the combination of these three phenotypes is a very strong indication of a mutation in a centrosomal protein (see below).

6.5 Mechanosensory Mutants

Several centrosome and basal body-specific proteins were identified by a screen for mechanosensory mutants. This screen led to the identification of the basal body protein unc (Baker et al. 2004), the mecB allele of ana1 (Blachon et al. 2009), and the mecD allele of asl (Blachon et al. 2008).

A screen for mechanosensory mutants looks for adult lethal mutations that have no or an abnormal mechanoreceptor current. Since the mechanosensory apparatus is located in cilia, loss of the centrosome results in a loss of the mechanoreceptor current. Therefore, as in mutants of genes involved in various aspects of mechanotransduction (Chung et al. 2001; Walker et al. 2000), centriolar mutants are adult lethal and have no mechanoreceptor current (Fig. 1.8).

6.6 PCL Mutants

The PCL is a centriolar structure resembling an early intermediate in centriole biogenesis in its composition and in the genetic pathway that underlies its formation (Blachon et al. 2009). Since the PCL is similar to an early centriolar intermediate, it was postulated that it should be possible to identify mutants of genes required in early centriole formation by screening for mutants that do not form a normal PCL.

Use of this approach led to the discovery of a mutant in the Drosophila ortholog of Poc1. Since it is thought that the PCL is related to male fertility, male sterile mutants were screened for PCL defects. PCL presence was scored using a centriolar protein that labels the PCL strongly (Ana1). Poc1 is a conserved centrosomal protein found in protists and mammals and its localization suggests that it plays a role in early centriole formation (Blachon et al. 2009; Keller et al. 2009; Keller et al. 2005; Pearson et al. 2009).

7 Identification of Centrosomal Complexes

Drosophila embryos are a rich source for centrosomal protein complexes (Gopalakrishnan et al. in press; Kellogg and Alberts 1992; Moritz et al. 1995). Originally, this system was used to isolate CP190 and CP60 complexes, which are nuclear and centrosomal proteins (Kellogg and Alberts 1992; Kellogg et al. 1995). Later, this system was used to purify γ-TuRC and γ-TuSC complexes, which are found in the cytosol and reside in the PCM (Moritz et al. 1998; Oegema et al. 1999). More recently, drosophila embryos were used to isolate complexes of specific centrosomal proteins such Sas-6 (Gopalakrishnan et al. 2010) and Sas-4 (Gopalakrishnan et al. in press). Sas-6 forms homomers that are hypothesized to be the building block of the centriole central tubule (Gopalakrishnan et al. 2010; Kitagawa et al. 2011; van Breugel et al. 2011) (Fig. 1.13).

Fig. 1.13
figure 13

Sas-6 homomers as the building block of centriole central tubule. Sas-6 forms homomers in the cytosol. These tetramers are hypothesized to polymerize and form the central tubule of the centriole (Gopalakrishnan et al. 2010). (Illustrated by Iwasa, Janet Haru)

Full size image

Sas-4 forms a complex (named S-CAP) with the centrosomal proteins CNN, Asl, and D-PLP (Gopalakrishnan et al. 2011) (Fig. 1.14). Interestingly, mutations in the orthologs of these proteins results in microcephaly, a developmental disorder where brain size is severely reduced (Al-Dosari et al. 2010; Bond et al. 2005; Kalay et al. 2011; Thornton and Woods 2009). The finding that microcephaly linked proteins form a common complex may explain why mutations in any of these lead to the same disorder. The S-CAP complex also contains CP190 and tubulin (Gopalakrishnan et al. 2011).

Fig. 1.14
figure 14

S-CAP complex. a Sas-4 forms a complex known as S-CAP together with Cnn, Asl, and D-PLP. This complex also contains CP190 and Tubulin. b S-CAP complexes are tethered to the centrosome via Sas-4 and contribute to the formation of the PCM (Gopalakrishnan et al. in press) (The fig is modified from Fig. 1.7 (Gopalakrishnan et al. 2011)

Full size image

Drosophila embryos are also a rich source for centrosomes. These centrosomes can nucleate microtubules in a cell-free system, allowing the study of the mechanisms of astral microtubules formation (Moritz et al. 1995). The function of centrosomal complexes can be studied further by stripping the centrosome from its PCM using high salt and then adding embryo extract and purified γ-TuRC complexes (Moritz et al. 1998). Stripping the centrosome is also useful to identify the mechanism of recruiting other centrosomal complexes and was used to show that Sas-4 is the S-CAP component that tethers the complex to the centrosome (Gopalakrishnan et al. 2011).