A Look at Drosophila Pattern Formation

Researchers interested in gene expression studies adopt one of two approaches. They can either examine the expression of a given gene in a population of cells in aggregate, or they can study the gene on a cell-by-cell basis in situ. The advantage of the former approach is its simplicity: It is generally easy to prepare RNA or protein from a given tissue sample and to probe it for the gene or protein of interest. But there are several disadvantages associated with the population approach. First o

Sep 3, 2001
Jeffrey Perkel
Researchers interested in gene expression studies adopt one of two approaches. They can either examine the expression of a given gene in a population of cells in aggregate, or they can study the gene on a cell-by-cell basis in situ. The advantage of the former approach is its simplicity: It is generally easy to prepare RNA or protein from a given tissue sample and to probe it for the gene or protein of interest. But there are several disadvantages associated with the population approach. First of all, it is difficult, if not impossible, to obtain spatial information--that is, determining in which cells, and where within the cell, the gene is being expressed. Population studies are complicated if the cell population of interest is a minority of the total cell population, a situation that would lead to problematic signal-to-noise ratios. Finally, population studies tend to require relatively large numbers of cells. In contrast, in situ techniques provide spatial detail, can detect very small numbers of positive cells, and require relatively small cell numbers overall. However, these technologies also tend to be both more technically challenging and laborious.

In situ macromolecular analyses typically have five steps. The tissues must be fixed, permeabilized, and probed for the molecule of interest. RNA species can be detected using tagged DNA or RNA probes. Protein species, in contrast, are typically detected using a tagged antibody. The samples are then mounted onto slides and analyzed microscopically. In situ methods have advanced through improvements in microscopy technologies, such as confocal and deconvolution microscopy. Traditionally, microscopic analysis of thick tissues could only be accomplished using sectioning of paraffin-embedded samples. However, confocal and deconvolution techniques allow the user to make optical sections through a thick sample, improving the resolution and removing fluorescent background from sections above and below the focal plane without introducing the artifacts that can occur with manual tissue sectioning. These techniques also allow the user to reconstruct the original three-dimensional image, so that staining can be viewed in the context of the entire tissue sample.

In this article, The Scientist examines a process that has been meticulously dissected--the embryonic pattern formation in the fruit fly, Drosophila melanogaster.1,2 This work represents an excellent example of the type of data that can be obtained using both in situ mRNA hybridization (ISH) and immunological detection protocols. Indeed, owing to technical difficulties, most of this data could not have been collected by any other method.

Researchers studying development have long pondered how the various body axes (anterior-posterior and dorso-ventral) are formed. Drosophila is an excellent model for this type of study, because the organism can be manipulated easily in the lab, has hundreds of well-characterized mutations, and has been highly characterized developmentally.1 In fact, fly development is so well documented that it is possible to know at exactly which stage of development a given embryo is at based on its morphological features. Recently, researchers published the sequence of the entire Drosophila genome, adding greatly to the utility of this organism as a model system.3

The developing Drosophila embryo is typically divided into four systems: anterior, posterior, terminal, and dorsoventral.1,2 The anterior system defines the head and thorax; the posterior system defines the abdominal segments; the terminal system defines the structures at the unsegmented ends of the egg; and the dorsoventral system defines the top-to-bottom axis (in humans, the dorsoventral pattern defines our backs and fronts, respectively).

Pattern Formation Explained

Pattern formation in the embryo results from coordinated and temporally regulated expression of both maternal and embryonic genes. During Drosophila oogenesis, the single germ cell divides four times, producing a structure containing 15 nurse cells and one egg, all interconnected via a series of cell-to-cell channels called ring canals. Surrounding this structure, somatic follicle cells nurture and support the egg chamber.1,2 Maternal genes play a role in development when the nurse cells extrude mRNA species into the egg itself. One well-known example of this involves the gene bicoid, the primary anterior system determinant. The bicoid gene encodes a transcription factor, responsible for activating zygotic genes, including hunchback, another transcriptional factor. ISH analyses have shown that nurse cells surrounding the oocyte deposit bicoid mRNA at what will become the anterior end of the fly.1,2 The bicoid mRNA is translated into protein in the embryo, forming a concentration gradient that is highest at the point of mRNA deposition and that diminishes towards the posterior, which can be detected by immunological techniques. Thus, bicoid encodes a morphogen, a protein that determines the local pattern of differentiation based upon its local concentration.1,2

The importance of this placement is demonstrated by the fact that, if bicoid mRNA is experimentally injected at another location within a bicoid mutant embryo, the head structures of the resultant fly will develop from that site.1,2 Furthermore, if higher concentrations of bicoid mRNA are placed in the nascent embryo, the size of the developing anterior structures will proceed towards the posterior at the expense of posterior structures. Thus, bicoid is said to be instructive for anterior development, as it is directly responsible for the anterior determination pathway.

Although it is clear that maternal nurse cells insert bicoid mRNA into the oocyte, researchers nevertheless wondered how the transcript could be so tightly localized to the extreme anterior end of this cell. Why wouldn't the mRNA, once deposited, become evenly distributed around the oocyte? Recently, Byeong-Jik Cha, Birgit S. Koppetsch, and William E. Theurkauf, at the University of Massachusetts Medical School in Worcester, Mass., published a paper in Cell that sheds some light on this question.4

Earlier publications had suggested a model in which developmental determinants are concentrated at the poles via microtubule-based motors, presumably by way of a polarized network of microtubule filaments.5 The basic idea, says Theurkauf, is that minus end-directed microtubule motors would carry the bicoid transcripts towards the anterior end, whereas positive end-directed motors would transport posterior determinants towards the other end of the oocyte. Such a model would require a highly polarized network of fibers, in which the minus ends of the filaments concentrated at the anterior, extending towards the posterior; however, no such network of filaments could be visualized.

Theurkauf and his colleagues were studying bicoid mRNA migration by injecting in vitro transcribed RNA probes labeled with either fluorescein isothiocynate (FITC) or Alexa Fluor 546 into either the nurse cells or the oocyte directly. They then visualized the movement of the labeled transcripts using time-lapse laser scanning confocal microscopy. About a year ago, they observed that if naked bicoid mRNA was injected directly into oocytes, it would migrate to the nearest cortical surface. Theurkauf notes that, since the nurse cells contact the anterior side of the oocyte, they assumed they now understood how the RNA was concentrated at the anterior end: As the nurse cells pumped bicoid mRNA into the cells, it would concentrate at the anterior pole, the nearest surface. However, when conducting what they believed was a routine control for their experiments, they found that, if the RNA was first injected into nurse cells prior to injection into the oocyte, it migrated not to the nearest cortical surface, but specifically toward the anterior pole. In a series of experiments conducted to understand their results, these researchers determined that, in the nurse cells surrounding the oocyte, bicoid mRNA forms a complex with the protein Exuperantia, in a microtubule-dependent manner. This complex then translocates into the oocyte and migrates to the anterior pole, also in a microtubule-dependent process.

Three elements critical to this process are microtubules, Exuperantia protein, and the bicoid 3'-untranslated region (the bicoid localization element 1, BLE1). In the absence of microtubules, bicoid mRNA fails both to recruit Exuperantia in the nurse cells, and to migrate into the oocyte. In Exuperantia mutant nurse cells, complexes form and migrate to the oocyte, but fail to translocate to the anterior pole. On the other hand, if bicoid mRNA is injected into wild type nurse cells, and transferred into Exuperantia mutant oocytes, then the RNA can properly migrate to the anterior pole, suggesting that Exuperantia is required in the nurse cells, but not in the oocyte. Finally, BLE1-deleted bicoid mRNA injected into nurse cells failed both to efficiently pass through the ring canals into the oocyte and to accumulate at the anterior pole of the oocyte. The authors suggest that a fourth element, a hypothesized nurse cell polarity factor, is also important in this process, and add that this activity could be a microtubule motor that allows the bicoid mRNA/Exuperantia complex to preferentially utilize microtubules originating at the anterior cortex in the oocyte.

There are, of course, other Drosophila morphogens, but not all of them are RNAs. In 1989, researchers in the lab of Christiane Nüsslein-Volhard at the Max-Plank Institute in Tübingen, Germany, demonstrated that a gradient of nuclear localization of the maternal Dorsal protein defines dorsoventral patterning in the Drosophila embryo.6 There is, however, no gradient of dorsal mRNA localization. Another maternal gene, Toll, encodes a transmembrane receptor, which is distributed evenly around the embryo. On the ventral (bottom) side of the embryo, spätzle, the Toll ligand, binds to and activates Toll, initiating a signal transduction cascade that results in the nuclear translocation of the Dorsal protein on the ventral side of the embryo only. Immunodetection of Dorsal protein in transverse sections of developing Drosophila embryos shows strongly labeled nuclei on the ventral side of the embryo, and completely unlabeled nuclei on the dorsal side. Along the dorsoventral midline of the organism, nuclear and cytoplasmic staining is mixed. Careful analysis of a series of dorsoventral pattern mutants demonstrated that the amount of nuclear Dorsal protein ultimately defines how "ventral" a region of the embryo will become.

In Situ Detection of Macromolecules

Researchers usually study mRNA localization in embryos using ISH.7,8 Fundamentally, there is no difference between ISH and other hybridization techniques such as Southern and Northern hybridizations. However, there are several steps that are required for ISH of Drosophila tissues that distinguish this technique from nylon and nitrocellulose membrane-based ones. First, tissues must be treated with either formaldehyde or paraformaldehyde, in order to fix the macromolecules in place. Additionally, Drosophila embryos are surrounded by two tissues that can interfere with the delivery of macromolecular probes into the tissue--the chorion and the vitelline membrane. Scientists remove the chorion using a solution of 50 percent commercial bleach (about five percent sodium hypochlorite), and use methanol and heptane to remove the vitelline membrane.

Probes can be composed of either DNA or RNA, and can use either a fluorescent label or a hapten (e.g., biotin or digoxigenin) moiety (nucleic acid labeling strategies were recently reviewed in The Scientist).9 Researchers detect hybridization events either via fluorescence microscopy or by detection of the hapten moiety. For example, digoxigenin-labeled probes can be detected using antibodies against digoxigenin that are coupled to alkaline phosphatase (AP), followed by incubation with 5-bromo-4-chloro-3-indoylphosphate and Nitro-blue tetrazolium (BCIP/NBT).

Immunodetection of proteins also requires removal of the chorion and vitelline membranes, as well as tissue fixation. Scientists use antibodies directed against the particular protein of interest, usually followed by incubation with a secondary antibody coupled to a detectable marker (either a fluorescent dye or an enzyme, such as AP or horseradish peroxidase [HRP]). Some researchers combine RNA and protein detection techniques in a single sample. In this case, the scientists stain for the protein of interest prior to ISH. In situ hybridization and immu- nodetection protocols can be found on the Web. Two sites offering a wide range of these and other protocols are LabVelocity (www.labvelocity.com) and Protocol Online (www.protocol-online.net).

Newly hatched fly researchers can also find extensive Drosophila resources, including protocols, on the Web. The Drosophila virtual library contains a useful list of protocols (www.ceolas.org/VL/fly/index.html), including information on animal husbandry and staining. Other resources include The Interactive Fly (sdb.bio.purdue.edu/fly/aimain/1aahome.htm); FlyBase, a genome resource (flybase.bio.indiana.edu); JFly, a data depository for Drosophila researchers (jfly.nibb.ac.jp); and FlyBrain, an online atlas of the Drosophila nervous system (flybrain.neurobio.arizona.edu). This listing is only a small sampling of the information available online; a wealth of other sites exist as well, easily found by following "additional resources" links from these and other related pages.

It is difficult to overstate the role played by in situ technologies in the dissection of pattern formation pathways in the Drosophila embryo. The images produced in these experiments, by demonstrating the concentration and relative positioning of maternal and zygotic factors that lead to proper development, can generate insights into the functions of genes that were previously not understood. In order to accomplish the same work using population analysis of protein and RNA, scientists would have to precisely excise regions of interest from the tissues, and then extract the macromolecules from them. However, without the in situ data available, it would not have been possible to know what regions to examine in the first place. Thus, without in situ detection technologies, fly researchers might still be very much in the dark.

Jeffrey M. Perkel can be contacted at jperkel@the-scientist.com.
1. P.A. Lawrence, The Making of a Fly: The Genetics of Animal Design, Cambridge, Mass., Blackwell Scientific Publications, 1992.

2. D. St. Johnston, C. Nüsslein-Volhard, "The origin of pattern and polarity in the Drosophila embryo," Cell, 68:201-19, 1992.

3. M.D. Adams et al., "The genome sequence of Drosophila melanogaster," Science, 287:2185-95, March 24, 2000.

4. B.-J. Cha et al., "In vivo analysis of Drosophila bicoid mRNA localization reveals a novel microtubule-dependent axis specification pathway," Cell, 106:35-46, July 13, 2001.

5. A. Bashirullah et al, "RNA localization in development," Annual Review of Biochemistry, 67:335-94, 1998.

6. S. Roth et al., "A gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosophila embryo," Cell, 59:1189-1202, 1989.

7. B. Sinclair, "Deep into that darkness peering: Fluorescence in situ hybridization," The Scientist, 13[17]:17, Aug. 30, 1999.

8. Roche Molecular Biochemicals, Nonradioactive In Situ Hybridization Application Manual, biochem.boehringer-mannheim.com/prod_inf/manuals/insitu/insi_toc.htm.

9. D. Stull, "A feast of fluorescence," The Scientist, 15[10]:20, May 14, 2001.

Selected Drosophila Web Resources
Berkeley Drosophila Genome Project

FlyMove - Exploring Drosophila Embryogenesis

FlyTrap - HTML-based Gene Expression Database

FlyView - A Drosophila Image Database