I. INTRODUCTION: FROM GENES TO INTEGRATIVE PHYSIOLOGY

Top Previous Next References

Physiology is defined as the function of a living organism and its parts. The enormous breadth of physiological research, which ranges from the study of single molecules to whole organisms, reflects the inherent complexity of life. The stunning progress in molecular biology that occurred during the last 50 years drove a powerful reductionist approach in physiological research. That same progress, however, now forms the foundation for the next revolution in the field. This revolution will be focused on integrative physiology, which seeks to understand multicomponent processes and the underlying pathways of information flow from an organism's "parts" up through increasingly complex levels of organization. Identification of the genes and proteins involved in specific physiological processes will lead ultimately to the integrated molecular understanding of underlying signaling and metabolic pathways, macromolecular protein assemblies, and functional interactions of organelles, cells, tissues, and organs.

This review has two main goals. The first goal is to describe the experimental utility of nonmammalian model organisms for investigating basic physiological problems. What is a model organism? The physiologist and Nobel laureate August Krogh believed that there is an ideal species or "model system" in which almost every biological problem can be studied most readily, a belief that is often referred to as the "Krogh Principle" (175, 176). Krogh would define a model organism as any organism that provides experimental characteristics that are optimal for the study of a physiological process.

In the postgenome sequencing era, the term model organism has taken on a more explicit meaning. Model organisms are "simpler," genomically defined organisms that are easy to grow in the laboratory, produce large numbers of offspring, and have relatively short life cycles making them suitable for detailed genetic analyses. These characteristics allow more rapid, efficient, and economical manipulation, and hence understanding, of gene function than what is possible in "complex" organisms.

Despite the narrowing of the definition of model organism, the Krogh Principle applies, perhaps more than ever, to the field of integrative physiology. If one wishes to define the genes and genetic pathways that give rise to complex physiological processes, doing so in a simple model system that is genetically tractable, and where it is relatively easy to manipulate gene expression makes sound experimental sense. Organisms such as Saccharomyces, Drosophila, and even the plant Arabidopsis are powerful experimental systems for addressing physiological problems common to all eukaryotes. The nematode Caenorhabditis elegans is an exquisite example of the experimental advantages provided by genetically tractable model organisms. This review provides a detailed overview of C. elegans biology and the experimental tools, resources, and strategies available for its study.

Ion channels and transporters play fundamental roles in organelle, cell, organ, and whole organism function. These roles range from regulation of organelle pH, cell volume, and membrane excitability to control of systemic salt and water balance and behavior. The second goal of this review is to describe how forward and reverse genetic approaches and direct behavioral and physiological measurements in C. elegans have generated novel insights into the integrative physiology of ion channels and transporters. I will also describe how insights from C. elegans have provided new understanding of the physiology of membrane transport processes in mammals.

	II. THE NEMATODE CAENORHABDITIS ELEGANS

Top Previous Next References

The establishment of C. elegans as a model system for fundamental biological research began in 1963 with the efforts of Sydney Brenner, a molecular biologist then at the Medical Research Council (MRC) Laboratory of Molecular Biology in Cambridge, United Kingdom. In a letter to Max Perutz, the head of the laboratory, Brenner (37) stated that "it is widely realized that nearly all the 'classical' problems of molecular biology have either been solved or will be solved in the next decade" and that the "future of molecular biology lies in the extension of research to other fields of biology, notably development and the nervous system."

The extraordinary successes that had been achieved at that time in defining the molecular bases of biological processes in bacteria suggested to Brenner that a similar approach would work in more complex organisms. He told Perutz that he wanted to "define the unitary steps of development using the techniques of genetic analysis." To tackle this problem, Brenner felt that he would need to identify a metazoan animal that he could "microbiologize" and handle in much the same way as bacteria and viruses. Brenner concluded his letter by stating that he wanted to "tame a small metazoan organism to study development directly."

Despite criticism from other molecular biologists that his approach was too "biological," Brenner was asked to submit a formal proposal to the MRC. He did so in October 1963, proposing to study the genetics of differentiation in the nematode Caenorhabditis briggsae.¹ The possible utility of nematodes in genetic research had first been noted by Dougherty and Calhoun (71). In Brenner's opinion, Caenorhabditis provided numerous experimental advantages for identifying genes involved in development. The animal has a short life cycle, produces large numbers of offspring by sexual reproduction, and can be cultured easily in the laboratory. Sexual reproduction occurs by self-fertilization in hermaphrodite worms or by mating with males, which makes Caenorhabditis exceptionally useful for genetic studies. Self-fertilization allows homozygous worms to breed true and greatly facilitates the isolation and maintenance of mutant strains. It is also a handy feature if mutant animals are paralyzed or uncoordinated since reproduction does not require movement to find and mate with a male. Mating with males, however, is essential for moving mutations between strains. Finally, Caenorhabditis is a highly differentiated animal but is comprised of <1,000 somatic cells and therefore provides a tractable system for studies of metazoan cellular function, development, and differentiation.

Brenner concluded his proposal with the following statement, "To start with we propose to identify every cell in the worm and trace lineages. We shall also investigate the constancy of development and study its control by looking for mutants." The statement is audacious and will surely amuse anyone who has struggled through the National Institutes of Health funding system. Nevertheless, it was Brenner's audacity and the good judgement of the MRC that established C. elegans as a model organism.

C. elegans researchers have accomplished Brenner's goal of tracing the lineage of every nematode cell and identifying genes responsible for development. In 2002, Sydney Brenner shared the Nobel Prize in Physiology or Medicine with John Sulston and H. Robert Horvitz for their discovery of genes in C. elegans that regulate organ development and programmed cell death. However, as discussed below, the extraordinary experimental power of C. elegans has been exploited well beyond studies of developmental biology. This tiny animal has provided and undoubtedly will continue to provide important insights into fundamental biological problems such as ageing, RNA-mediated gene silencing, cell cycle control, sensory physiology, and synaptic transmission.

B.  C. elegans Biology

1.  Natural history and life cycle

C. elegans is a member of the phylum Nematoda. Nematodes, or roundworms, are some of the most numerous and widespread of all animals and are found in virtually all habitats. The phylum contains free-living species as well as species that parasitize plants and other animals. In addition to its extraordinary utility as a model for basic biological research, detailed understanding of all aspects of C. elegans function has major economic and medical implications. Approximately one-half the world's human population is infected with parasitic nematodes (205), and parasitic plant nematodes cause an estimated $80 billion in crop damage annually (287).

Nematodes range in size from <1 mm to over 35 cm in length. C. elegans is a free-living soil nematode ~1 mm long. The life strategy of C. elegans is well adapted for survival in soil environments where food and water availability, temperature, populations of predators, and many other variables can change constantly and dramatically. It is a voracious feeder that, as noted by Riddle et al. (273), consumes "anything that fits in its mouth." To outgrow competitors, C. elegans produces large numbers of offspring and rapidly depletes local food resources.

Adult C. elegans are predominantly hermaphroditic with males making up ~0.1% of wild-type populations. Self-fertilized hermaphrodites produce ~300 offspring, whereas male-fertilized hermaphrodites can produce over 1,000 progeny.

Under optimal laboratory conditions, the average life span of C. elegans is 2-3 wk. The life cycle is rapid (Fig. 1). At 25 °C, embryogenesis, the period from fertilization to hatching, occurs in ~14 h. Postembryonic development occurs in four larval stages (L1-L4) that last a total of ~35 h.

View larger version (22K): [in this window] [in a new window]   Fig. 1. Life cycle of Caenorhabditis elegans grown at 25°C on agar plates seeded with Escherichia coli. Eggs are laid ~5 h after fertilization, and hatching occurs ~9 h later. [From Jorgensen and Mango (153), with permission from Nature Reviews Genetics, copyright 2002 Macmillan Magazines Ltd.]

When food supply is limited, dauer larvae form after the second larval molt. Dauer larvae do not feed and have structural, metabolic, and behavioral adaptations that increase life span up to 10 times and aid in the dispersal of the animal to new habitats. Once food becomes available, dauer larvae feed and continue development to the adult stage (243, 297).

The standard C. elegans laboratory strain is Bristol N2. Other strains are also used and offer certain experimental advantages. For example, the Bergerac strain exhibits a high rate of spontaneous mutation due to the activity and high copy number of the Tc1 transposon (see sect. IIC3D) (220). The Hawaiian strain CB4856 possesses a uniformly high density of single nucleotide polymorphisms that greatly facilitate genetic mapping (349).

Culture of C. elegans in the laboratory is simple and relatively inexpensive (187). Animals are typically grown in petri dishes on agar seeded with a lawn of E. coli as a food source. C. elegans can also be grown in mass quantities using liquid culture strategies and fermentor-like devices. Worm stocks are stored frozen in liquid nitrogen indefinitely with good viability. The ability to store C. elegans frozen dramatically simplifies culture strategies and reduces costs associated with handling and maintaining wild-type and mutant worm strains.

Like all nematodes, C. elegans has an unsegmented, cylindrical body that tapers at both ends. The body wall consists of tough collagenous cuticle underlain by hypodermis, muscles, and nerves. A fluid-filled body cavity or pseudocoel separates the body wall from internal organs. Body shape is maintained by hydrostatic pressure in the pseudocoel (Fig. 2).

View larger version (22K): [in this window] [in a new window]   Fig. 2. Schematic diagrams showing anatomical features of an adult C. elegans hermaphrodite (top) and male (bottom). [From Sulston and Horvitz (307), with permission from Academic Press.]

Newly hatched L1 larvae have 558 cells. Additional divisions of somatic blast cells occur during the four larval stages eventually giving rise to 959 somatic cells in mature adult hermaphrodites and 1,031 in adult males. The lineage of somatic cells in C. elegans is largely invariant. This invariance combined with the ability to visualize by differential interference contrast (DIC) microscopy cell division and development in living embryos, larvae and adult animals has made it possible to describe the fate map or cell lineage of the worm (307, 308).

Despite the small cell number, C. elegans exhibits an extensive degree of differentiation. Many physiological functions found in mammals have nematode analogs. This high degree of complexity and small total cell number provides a remarkably tractable experimental system for studies of differentiation, cell biology, and cell physiology.

A) SKIN. The worm "skin" or hypodermis is an epithelium that underlies the cuticle. Hypodermal cells secrete the cuticle, provide a substrate for cell and axon migration, and function as a storage site for lipids and other molecules (126, 347). The hypodermis is also involved in phagocytosis of apoptotic cells (275, 308) and may have an osmoregulatory function (248). Certain types of hypodermal cells function as blast cells and give rise postembryonically to new cell types such as neurons (308). Most hypodermal cells are multinucleate.

B) MUSCLE. C. elegans is a particularly powerful model for defining muscle physiology, structure, molecular biology, and development (reviewed in Refs. 219, 343). The worm has a number of readily observable characteristics that allow rapid screening for mutations in genes involved in muscle function. Even mutations that cause severe locomotion defects can be studied readily and in detail because hermaphrodites self-fertilize and locomotion is unnecessary for reproduction.

C. elegans possesses both striated and nonstriated muscles. Striated body wall muscles are the most numerous muscle cell type. They are arranged in longitudinal bands along the body wall and are responsible for locomotion. Nonstriated muscles are associated with the pharynx, intestine, anus, and the hermaphrodite uterus, gonad sheath, and vulva. These muscles are responsible for pharyngeal pumping, defecation, ovulation and fertilization, and egg laying. Specialized nonstriated muscles are located in the tail of male worms and function in mating.

C) NERVOUS SYSTEM. The nervous system of adult hermaphrodites contains 302 neurons and 56 glial and support cells. Males have 381 neurons and 92 glial and support cells. White et al. (348) have reconstructed and mapped the connectivity of the entire hermaphrodite nervous system using serial electron microscopy. Most of the differences between the male and hermaphrodite nervous system are found in the male tail, which plays an important role in mating. Partial reconstruction of the tail nervous system has been performed (306).

The nervous system of C. elegans is generally thought of as being divided into the pharyngeal and central nervous systems. Twenty neurons innervate and regulate the activity of the pharynx. These neurons are connected to the central nervous system by two interneurons.

Neuronal processes of the central nervous system are oriented as bundles that run along the hypodermis in longitudinal cords or circumferential commissures. The ventral and dorsal nerve cords emanate from the circumpharyngeal nerve ring. Interneurons and motor neurons make up the ventral nerve cord. The dorsal nerve cord is formed largely of ventral cord motor neuron axons that enter via commissures. Sense organs that respond to chemicals, temperature, mechanical force, and osmolality are located primarily in the head and tail.

An important feature of the C. elegans nervous system is that only three neurons, CANL, CANR and M4, are required for viability under laboratory conditions. The CAN (excretory canal) neurons run along the excretory canals (see sect. IIB3D) and may play an important role in regulating systemic salt and water balance. M4 is a pharyngeal motor neuron that controls isthmus peristalsis and thus controls feeding (12). The nonessential nature of most neurons for viability provides an enormous advantage for mutagenesis studies of nervous system function.

D) KIDNEY. The worm "kidney" consists of three cell types: the excretory cell, the duct cell, and the pore cell (230). Destruction of any of these cells by laser ablation causes the animal to swell with fluid and die (231).

The excretory cell is a large, H-shaped cell that sends out processes both anteriorly and posteriorly from the cell body (e.g., see Fig. 5). A fluid-filled excretory canal is surrounded by the cell cytoplasm. The basal pole of the cell faces the pseudocoel while the apical membrane faces the excretory canal lumen. Gap junctions connect the excretory cell to the hypodermis, suggesting an interaction between the two cell types important for whole animal osmoregulation.

An excretory duct connects the excretory canal to the outside surface of the worm. The duct is formed from cuticle that is continuous with the animal's exoskeleton. A duct cell surrounds the upper two-thirds of the duct and a pore cell surrounds the lower one-third.

The excretory cell is a single-cell "epithelium" that appears to secrete salt, water, and waste products into the excretory canal. Duct cells may also play an important role in solute and water transport. The apical surface area of the duct cell is greatly amplified by extensive invaginations, and the cytoplasm is filled with mitochondria. Nelson et al. (230) have suggested that the duct cell may be involved in selective solute reabsorption. If this is the case, the nematode excretory and duct cells are analogous to the acini and ducts of mammalian secretory epithelia such as the salivary gland, sweat glands, and pancreas. The excretory cell may be a particularly valuable model for defining molecular mechanisms of stimulus-secretion coupling (reviewed in Refs. 130, 162) and ion channel and transporter trafficking and polarization.

E) DIGESTIVE TRACT. The digestive tract of C. elegans consists of a pharynx, intestine, and rectum. C. elegans is a filter feeder, and the pharynx is a muscular organ that pumps food into the pharyngeal lumen, grinds it up, and then moves it into the intestine. The pharynx is formed from muscle cells, neurons, epithelial cells, and gland cells (3).

A pharyngeal-intestinal valve connects the pharynx to the intestine. Twenty epithelial cells with extensive apical microvilli form the main body of the intestine (186). Intestinal epithelial cells secrete digestive enzymes and absorb nutrients. The intestine functions as one of the major storage organs in the body. Intestinal cells are filled with numerous granules that likely contain lipids, proteins, and carbohydrates (173). The intestine also produces yolk proteins and secretes them into the pseudocoel where they are then taken up by oocytes (111, 116, 165).

The rectum is made up of five epithelial cells and is connected to the intestine via the intestinal-rectal valve. A sphincter muscle wraps around the valve and controls its opening. The sphincter muscle, an anal depressor muscle, and two muscle cells that wrap around the posterior intestine control defecation.

F) GONAD. The gonad of adult hermaphrodites consists of two identical U-shaped arms connected via spermatheca to a common uterus (116, 131, 140) (e.g., see Fig. 8A). The gonad arms are surrounded by thin epithelial cells termed sheath cells. The distal portion of each arm contains germline nuclei that give rise to sperm and oocytes. Each nucleus is surrounded by cytoplasm and an incompletely formed plasma membrane. The nuclei in turn are arranged in a monolayer that surrounds a central core of cytoplasm (e.g., see Fig. 8A).

The germline nuclei population is maintained by mitosis occurring in the distal-most tip of each gonad arm. As germ cells progress down the arm, they exit the mitotic cell cycle and undergo meiosis. Formation of an intact germ cell plasma membrane, a process termed cellularization, occurs in the proximal arm of the gonad. During the fourth larval stage, germ cells differentiate into ~150 spermatids/gonad arm. These spermatids develop into mature sperm that are stored in the spermatheca.

In adult worms, germ cells differentiate into oocytes. Oocytes are arranged in the proximal gonad arm in a single-file fashion. Once formed, oocytes undergo oogenesis, which is a period of intense biosynthetic activity and rapid and massive oocyte growth.

Oocytes remain in diakinesis of prophase I until they reach the most proximal position in the gonad arm. During the late stage of oogenesis, an oocyte located immediately adjacent to the spermatheca undergoes meiotic maturation (e.g., see Fig. 8, A and B). Within 5-6 min after maturation is initiated, the oocyte is ovulated into the spermatheca where it is fertilized. Completion of the meiotic divisions occurs in the uterus and is followed by embryogenesis. In a mature hermaphrodite, ovulation occurs once every 20-40 min. Unmated hermaphrodites produce ~300 progeny over a 3-day period when grown under standard laboratory conditions (116, 131, 341).

The testis, seminal vesicle, and vas deferens form the male gonad (131, 168). The testis is U-shaped and is formed from two distal tip cells that surround the germ cells. As in the hermaphrodite gonad, germ cell nuclei at the distal end of the testis undergo mitosis. Germ nuclei farther away enter prophase of meiosis I. Spermatogenesis begins in the proximal end of the testis.

The seminal vesicle is formed by 20 secretory cells. Spermatocytes enter the vesicle and undergo two meiotic divisions to form mature spermatozoa. Spermatozoa are stored in the seminal vesicle. The vas deferens is comprised of 30 cells and functions as a valve that controls release of sperm during ejaculation.

C.  Tools and Experimental Strategies for Studying Ion Channel and Transporter Biology in C. elegans

1.  Cell physiology

A) TARGETED KILLING OF IDENTIFIED CELL TYPES. A powerful way to assess the physiological role of a specific nematode cell type is to destroy the cell and characterize the effect on developmental events and whole animal phenotype. Laser ablation or microsurgery has been used extensively to identify cell function and cell-cell developmental interactions in C. elegans (18). Identified cell types are destroyed using a nitrogen-laser-pumped dye laser. The laser beam is focused onto the cell of interest through a high numerical aperture objective lens on a compound microscope equipped with DIC optics.

It is also possible to genetically target cells for killing. Maricq et al. (204) first used this approach to kill neurons expressing the glutamate receptor gene glr-1. The glr-1 promoter was fused to the mec-4(d) open reading frame and DNA transformation carried out using standard methods (see sect. IIC3B). mec-4(d) encodes a mutant DEG/ENaC cation channel that is thought to be constitutively active (72). Constitutive channel activation presumably results in excessive cation influx into the cell, which leads to cell swelling and necrotic cell death (115). Driscoll and co-workers (121) have demonstrated that ectopic expression of mec-4(d) induces degeneration of a wide variety of cell types. Targeted cell killing has also been carried out using human caspase-1 to induce apoptotic cell death (362) and diptheria toxin (87).

B) ELECTROPHYSIOLOGY OF DIFFERENTIATED SOMATIC CELLS. In vivo measurement of physiological processes such as ion channel and transporter activity in C. elegans is technically demanding and time consuming. Most somatic cells are quite small, and a tough pressurized cuticle surrounding the animal limits access to the cells for direct functional measurements.

To patch-clamp neurons, Lockery and co-workers (109, 198) developed the so-called "slit worm preparation" (Fig. 3A). Animals are immobilized by gluing them with cyanoacrylate glue onto a glass coverslip coated with a thin layer of agarose. Exposure of the neurons involves cutting a small slit in the worm with a glass dissecting needle. Groups of neurons containing several cell bodies emerge through the cut. Different types of neurons can be identified using cell-specific green fluorescent protein (GFP) reporters (see sect. IIC3C).

View larger version (72K): [in this window] [in a new window]   Fig. 3. Preparations used for electrophysiological measurements in C. elegans. A: schematic diagram and differential interference contrast (DIC) micrograph of "slit worm" preparation used for patch clamping of C. elegans neurons. Neuron expressing a green fluorescent protein (GFP) reporter is shown in green. (DIC image kindly provided by Dr. Shawn Lockery.) B: schematic diagram and fluorescence micrograph of "filleted worm" preparation used for patch clamping of C. elegans body wall muscles. GFP reporters are expressed in ventral nerve cord and body wall muscles. Scale bar is 50 µm. [From Richmond and Jorgensen (272), with permission from Nature Publishing Group.] C: schematic diagram and fluorescence micrograph of "filleted head" preparation used for patch clamping neurons. AVA interneurons are identified using GFP expressed under the control of the nmr-1 promoter. [From Brockie et al. (38), with permission from Elsevier Science.] D, top: schematic diagram and equivalent circuit of preparation for recording electropharyngeogram (EPG). [From Avery et al. (13), with permission from Academic Press.] Bottom: EPG from a wild-type worm. The first upward spike (E) indicates muscle depolarization that initiates contraction. Large downward spike (R) is muscle repolarization that precedes relaxation. Inhibitory postsynaptic potentials (IPSPs) generated by the M3 motor neuron occur when the muscle is depolarized. [From Dent et al. (69), with permission from Oxford University Press.]

Richmond and co-workers (271, 272) developed a "filleted worm" preparation to study body wall muscle electrophysiology (Fig. 3B). As with the slit worm preparation, animals are immobilized on agarose pads with cyanoacrylate glue. A lateral incision in the cuticle is then made to expose the body wall muscles.

A filleted preparation was described recently by Brockie et al. (38) that allows access to and partially preserves the spatial organization of neurons in the head. Briefly, worms are glued to Sylgard-coated coverslips and the cuticle along the length of the pharynx is slit with a fine glass needle. A flap of cuticle is then retracted and pinned to the Sylgard (Fig. 3C).

The C. elegans pharynx has proven to be an important model system for identifying the genetic basis of excitable cell function. Raizen and Avery (260) developed a simple extracellular recording method for monitoring electrical transients induced by changes in the membrane potential of pharyngeal muscles. Briefly, the head of a worm is sucked into a fluid-filled glass pipette and electrical connections between the pipette fluid and bath are made with silver/silver chloride electrodes (Fig. 3D). During an action potential, the basal membrane of the pharynx depolarizes. Excitation of the pharyngeal muscles causes current to flow across the apical membrane of the pharynx, out of the worm's mouth, and back into the pseudocoelom through the cuticle and hypodermis. These electrical transients are termed the electropharyngeogram (EPG). The EPG equivalent circuit and a typical EPG from a wild-type worm are shown in Figure 3D. Analysis of EPG in mutant animals has been particularly useful for defining the molecular bases of cell excitability (e.g., see sect. IIIG1).

The pharynx can also be isolated easily by cutting the animal transversely behind the pharyngeal bulb. After cutting, body wall muscles contract, retracting the cuticle covering the head and exposing the pharynx. The isolated pharynx can be impaled with glass microelectrodes (64, 65, 93, 245).

C) ELECTROPHYSIOLOGY OF GERM CELLS AND EARLY EMBRYOS. The first reported patch-clamp studies on C. elegans were carried out on spermatocytes and spermatids (200). Like those of other nematodes, C. elegans spermatozoa are nonflagellated ameboid cells (187a). Sperm cells can be isolated easily by cutting male worms with a needle close to the vas deferens.

Nematode oocytes are also relatively easy to isolate. Briefly, individual nematodes are cut behind the pharyngeal bulb using a 26-gauge needle. The extruded gonad is then detached by making another cut in front of the spermatheca. Isolated gonads contract and eject oocytes through the cut end. It is also possible to microdissect sheath cells away from the oocytes using glass micropipettes (279). Oocytes are readily patch clamped (54, 279) and can be loaded with voltage-, Ca²⁺-, and pH-sensitive fluorescent dyes (C. Boehmer, E. Rutledge, and K. Strange, unpublished observations).

Mass quantities of oocytes can be isolated from the fer (fertilization defective)-1 worm strain (8, 303). The fer-1 hermaphrodites produce normal sperm and are self-fertilizing at 18°C. However, at 25°C, the sperm do not develop properly and are incapable of fertilizing oocytes. Unfertilized oocytes are ovulated and accumulate in the uterus where they continue to undergo DNA replication and become polyploid. Induction of oocyte laying is triggered by treating worms with serotonin and levamisole (8) or by gentle sonication (303). Released oocytes are separated from adult worms and debris by filtering through mesh screens. In our experience, mass-isolated oocytes are difficult to patch clamp in the conventional whole cell mode (E. Rutledge and K. Strange, unpublished observations). However, they may be useful for other types of patch-clamp measurements as well as radioisotope flux studies.

Embryonic cells can be isolated easily and in large quantities. Adult nematodes are treated with an alkaline hypochlorite solution, which lyses the cuticle (187). Eggs released by this treatment are pelleted by centrifugation and washed in saline. Eggshells are removed by resuspending pelleted eggs in a chitinase solution (79). After digestion of the eggshell, early embryo cells are dissociated by passing the embryo suspension through a 23-gauge needle. As with oocytes, embryo cells can be patch clamped (54) and loaded with ion-sensitive fluorescent dyes (M. Christensen and K. Strange, unpublished observations).

D) CELL CULTURE. Because of technical hurdles, large-scale cell culture methods have not been widely exploited until recently for the study of C. elegans. Culturing differentiated cells from larvae and adult C. elegans is probably not technically feasible because of difficulties in removing the animal's cuticle and in dissociating cells. In contrast, embryonic cells can be isolated relatively easily (see sect. IIC1C). Early attempts at large-scale culture of C. elegans embryonic cells were described by Bloom (32). This pioneering work demonstrated that morphological differentiation occurred in these cells. However, Bloom noted significant problems with cell survival, attachment of cells to the growth substrate, cell differentiation, and reproducibility of the methods. Initial attempts to patch-clamp cultured cells were unsuccessful. Buechner et al. (41) have also reported that cultured C. elegans embryonic cells undergo morphological differentiation resembling that of neurons and muscle cells.

Recently, Christensen et al. (53) described methods that allow the robust, large-scale culture of C. elegans embryonic cells. Isolated embryonic cells differentiate within 24 h into the various cell types that form the newly hatched L1 larva. Expression of a number of cell-specific GFP reporters and molecular markers is similar to that observed in the intact animal (Fig. 4). Cultures can be generated from animals of almost any genetic background, and the cells can be readily patch clamped (53, 54) and loaded with ion-sensitive fluorescent dyes (A. Estevez and K. Strange, unpublished observations). Addition of double-stranded RNA (dsRNA) to the culture medium induces dramatic knockdown of targeted gene expression in cultured cells. Fluorescence-activated and magnetic-activated cell sorting can be used to enrich cultures for certain cell types. Enriched cultures can in turn be used for cell-specific biochemical, molecular, DNA microarray, and proteomic studies. C. elegans cell culture should be particularly valuable for physiologists interested in the molecular bases of ion channel and transporter function. When combined with other C. elegans tools, primary cell culture provides new opportunities for developing an integrated systems level understanding of eukaryotic ion channel and transporter biology.

View larger version (54K): [in this window] [in a new window]   Fig. 4. Micrographs of cultured C. elegans embryonic cells expressing cell-specific GFP reporters. A: myo-3::GFP expression in cultured body muscle cells. myo-3 encodes a myosin heavy chain isoform expressed in body muscles. The myo-3 reporter strain expresses two GFPs with peptide signals that target them to either the nucleus (arrows) or mitochondria (arrowheads). B: cultured mechanosensory neuron expressing mec-4::GFP. mec-4 encodes a degenerin-type ion channel subunit expressed largely in neurons that respond to gentle body touch. C: cultured cholinergic motor neurons expressing unc-4::GFP. unc-4 encodes a transcription factor. D: cultured neuron expressing opt-3::GFP. OPT-3 is a H+/oligopeptide transporter expressed in glutamatergic neurons. E: combined DIC and fluorescence micrograph of an unc-119::GFP-expressing neuron. unc-119 encodes a novel protein that is expressed in all neurons. F: combined DIC and fluorescence micrograph of an unc-4::GFP-expressing cholinergic motor neuron physically interacting with a body wall muscle cell. All scale bars are 10 µm. [From Christensen et al. (53), with permission from Elsevier Science.]

E) QUANTITATIVE MICROSCOPY. C. elegans is well-suited for quantitative, in vivo microscopy studies of ion channel and transporter activity. The embryo eggshell and cuticle of larvae and adults are transparent, making it possible to observe and quantitate cell biological events and physiological processes using bright-field and fluorescence microscopy. Differential interference contrast microscopy produces superb images of embryos, larvae, and adult animals. Larvae and adult worms are easily immobilized for imaging studies by anesthetizing with buffer containing 0.1% tricaine and 0.01% tetramisole or by gluing them with cyanoacrylate glue onto a glass coverslip or slide coated with a thin layer of agarose.

Proteins encoded by transgenes offer the potential to monitor membrane voltage and intracellular and intraorganelle ion levels in intact worms. For example, Ca²⁺-sensitive cameleons (216, 217) have been used to measure intracellular Ca²⁺ transients in pharyngeal muscle (160). GFP has been used as an optical sensor of membrane voltage changes induced by mechanical stimulation of C. elegans touch neurons in vivo (161). Other transgene-encoded cameleon proteins and GFPs may be useful for in vivo measurements of Cl (178), Ca²⁺ (16, 226), pH (170, 197), and signaling molecules that control ion transport processes such as cAMP (359), cGMP (134), and Ca²⁺-calmodulin (247). It is also possible to microinject oocytes (286) and intestinal cells (63) in vivo with ion-sensitive fluorescent dyes.

Fluorescence resonance energy transfer (FRET) (159) using the GFP color mutants CFP (cyan) and YFP (yellow) (214, 217) may be a powerful way to test for and further characterize protein-protein interactions in vivo in C. elegans that have been predicted by genetic screens and yeast two-hybrid and biochemical assays. This approach has been used recently to demonstrate a direct interaction between voltage-activated Ca²⁺ channels and calmodulin in transfected mammalian cells (85).

A) FORWARD GENETIC SCREENING. The development of C. elegans as an experimental system was driven largely by the relative ease of performing forward genetic screens for identification of the complement of genes responsible for observable phenotypes. The utility and power of genetic screening depends on the ability to assay a phenotype of interest. For a detailed and illuminating discussion of screening assays in C. elegans, the reader is referred to a recent review article by Jorgensen and Mango (153).

Once a screening assay is developed, animals are mutagenized, typically by the alkylating agent ethyl methane-sulfonate (EMS). Mutant animals are then isolated and the mutated gene identified by mapping, rescue, and cloning strategies (153). In addition to identifiying genes involved in a specific biological process, forward genetic screens can also provide important information on protein structure/function.

Mutant animals can be further mutagenized to produce double mutants. The phenotype of a double mutant could be suppressed, enhanced, or similar to the phenotype of animals with a single mutation. Mutations that enhance or suppress a phenotype may reside in genes distinct from the one mutated in the original screen. These extragenic mutations imply that the enhancer and suppressor genes interact with the first mutated gene. Genetic interactions indicate that gene products function in a common pathway. They also suggest the possibility of direct interactions between gene products (i.e., protein-protein interactions) (153). Ion channels, transporters, and associated regulatory machinery have been identified in genetic screens of various processes such as locomotion (4, 179, 206, 210, 261, 264), gonad development (345), mechanosensory (203, 314, 315) and osmotic avoidance behavior (58), chemotaxis and thermotaxis (225), resistance to fluoride ions (313), defecation (63, 64), egg laying (81, 152, 181), toxin and heavy metal sensitivity (39, 70, 92), programmed cell death (352), neuronal degeneration (322), and abnormal catecholamine levels (75).

B) THE GENOME. The C. elegans genome is diploid and consists of 97 × 10⁶ base pairs of DNA organized onto 5 autosomal chromosomes and one X sex chromosome. Recombinational mapping of mutant genes was first reported by Brenner (36) in 1974. In 1984, assembly of a clone-based physical map of the genome was begun (61). Work from numerous laboratories has subsequently led to the development of a detailed genome map with tight correlation between genetic and physical mapping data (60). The physical map of the worm genome is comprised of overlapping cosmid and yeast artificial chromosome (YAC) clones. These clones are freely available to the research community from the Sanger Center in Cambridge, United Kingdom (http://www.sanger.ac.uk/Projects/C_elegans/). The availability of these clones from a centralized resource has greatly facilitated genetic research in C. elegans. For example, identification of mutated genes is carried out by mapping the mutation to a genetic interval and then rescuing the mutant animal by microinjection of candidate cosmids (see sect. IIC3B).

The development and testing of technology required for the sequencing and understanding of the human genome was critically dependent on the analysis of genomes of less complex organisms such as C. elegans. Funding for sequencing of the worm genome was obtained in 1990 from the National Institutes of Health and the MRC in the United Kingdom. The sequencing project was a collaborative effort between the Sanger Institute in the United Kingdom and the Genome Sequencing Center at Washington University School of Medicine in St. Louis, Missouri. Sequencing of the worm genome was largely completed by the end of 1998 (318) and was the first sequence obtained for a multicellular organism. The genome contains ~20,000 predicted protein- and RNA-encoding genes (132). As of this writing, <10% of these genes have been studied experimentally.

To aid in gene identification, two major projects are ongoing. The C. elegans expressed sequence tag (EST) project undertaken by Kohara and co-workers (http://www.ddbj.nig.ac.jp/c-elegans/html/CE_INDEX.html) has tag sequenced over 60,000 cDNAs. These ESTs have identified more than 9,000 worm genes.

Vidal and co-workers (266) (http://worfdb.dfci.harvard.edu/) have undertaken an effort to clone all C. elegans open reading frames (ORFs), which they term the "ORFeome." ORFs are amplified from a cDNA library using specific primers and then sequenced to generate ORF sequence tags (OSTs). These OSTs support the existence of at least 17,300 C. elegans genes (266). To expedite functional analysis of the ORFs, they are cloned using the Gateway recombination cloning strategy (124, 336). This approach has the advantage of allowing directional cloning of PCR products into a reference vector and subsequent subcloning into various vector backgrounds without the use of restriction enzymes and ligases. As such, recombination cloning is amenable to automation and high-throughput analysis of gene function.

Gene identification in C. elegans has also been facilitated by development of the Intronerator (158) (http://www.cse.ucsc.edu/~kent/intronerator/), which is a series of internet tools for analyzing gene structure. Intronerator provides alignments of cDNAs with genomic sequence and a catalog of alternatively spliced genes.

Membrane transport physiologists are well acquainted with the use of hydrophobicity analysis for predicting transmembrane helices and, in theory, such an approach could be used for predicting membrane protein-encoding genes. The reliability of hydropathy methods, however, is limited (150, 169). Hidden Markov models (HMMs) with high accuracy for predicting transmembrane protein topology, have been developed recently (177, 325). HMMs are probabilistic models applicable to linear problems such as genome sequence analysis. Krogh et al. (177) (http://www.cbs.dtu.dk/services/TMHMM/) used an HMM to identify membrane-spanning proteins in the genomes of several organisms including C. elegans. Approximately 30% of the genes in the worm genome encode proteins with one or more predicted transmembrane helices.

Ian Paulsen and his group at the Institute for Genomic Research (TIGR) have analyzed the complement of predicted and known transport protein-encoding genes in the genomes of C. elegans and several other organisms (http://www.biology.ucsd.edu/~ipaulsen/transport/). These genes are classified into families according to the Transport Commission (TC) system (http://tcdb.ucsd.edu/tcdb/). The TC system is a scheme for classifying transport proteins based on functional and phylogenetic information and is analogous to the Enzyme Commission system for enzyme classification (281, 282). Paulsen's analysis has revealed that more than 750 genes or ~4% of the worm genome encode proteins thought to be involved in ATP-dependent, secondary active, and passive, channel-mediated transport processes.

B) CREATION OF TRANSGENIC ANIMALS. DNA transformation in C. elegans is relatively straightforward and was first reported by Stinchcomb et al. (300). Fire (88) and Mello et al. (213) have described methods for producing and maintaining transgenic worm lines. Briefly, transforming DNA is microinjected into the distal end of the hermaphrodite gonad. Heritable DNA transformation occurs by extrachromosomal transformation, nonhomologous integration, or homologous integration. Spontaneous homologous integration is extremely rare. Formation of multicopy extrachromosomal arrays is the most frequent way in which transforming DNA is inherited. Transformation by extrachromosomal arrays is often transient. Integration of transgenes and generation of stable transgenic lines is commonly carried out by gamma irradiation of transformed worms (212).

Recently, Praitis et al. (255) described the use of microparticle bombardment to create integrated transgenic lines in C. elegans. This approach is technically less demanding than microinjection and reportedly produces single- and low-copy chromosomal DNA insertions resulting in more reliable transgene expression. Low-copy integrated lines also exhibit expression of transgenes in cells of the germline. Multicopy extrachromosomal arrays are often not expressed in germ cells due to the activity of gene silencing processes (e.g., Ref. 157).

C) IDENTIFYING THE CELLULAR LOCALIZATION OF GENE PRODUCTS. Determining where and when a gene is expressed can provide important clues to its function. Gene expression patterns in C. elegans have been determined by in situ hybridization (30, 292, 311) and by creation of transgenic animals expressing lacZ (89) or GFP (49) reporter genes. Translational GFP reporters have been particularly useful for defining the cellular location and trafficking of proteins in C. elegans (e.g., Refs. 38, 78, 172). Examples of GFP reporter localization of ion channel and transporter proteins are shown in Figure 5. Databases of gene expression patterns in C. elegans are accessible online (http://nematode.lab.nig.ac.jp/db/index.html, http://129.11.204.86:591/default.htm, http://www.wormbase.org/).

View larger version (92K): [in this window] [in a new window]   Fig. 5. Examples of transgenic nematodes expressing cell-specific GFP reporters. A: fluorescence (left) and DIC (right) micrographs of a nematode expressing GFP in six dopaminergic neurons. GFP expression is driven by the promoter for the dopamine transporter (DAT). CEP and ADE dopaminergic neurons are indicated. Fluorescence micrograph is a 3-dimensional reconstruction generated by confocal microscopy. [From Nass et al. (227), copyright 2002 National Academy of Sciences USA.] B: fluorescence (left) and DIC (right) micrographs of a nematode expressing GFP in the H-shaped excretory cell (EC). GFP expression is driven by the clh-4b promoter. CLH-4b is a ClC chloride channel homolog. [From Nehrke et al. (229).]

D) GENE KNOCKOUT. Targeted gene inactivation by homologous recombination has not been feasible in C. elegans (252). Instead, the relative ease of culturing C. elegans in large numbers and the ability to store worms frozen has led to the development of so-called "target-selected gene inactivation methods." This approach involves inducing random deletion mutations in a population of worms. Mutagenized worms are subdivided into small pools, typically in 96-well culture plates. These pools are screened by PCR for a deletion mutation in a target gene. Once a pool containing the desired mutation is detected, it is subdivided and PCR screening is repeated until single worms carrying the deletion are isolated.

Zwaal et al. (365) were the first to use target-selected gene inactivation methods. Deletion mutations were screened for in cultures of mut (mutator)-2 worms in which endogenous transposons are actively "jumping" and inserting randomly into the genome (358). Bessereau et al. (28) have recently demonstrated that the Drosophila transposon Mos1 can be expressed and can function as a mutagen in the C. elegans genome. The use of heterologous transposons offers a number of advantages for forward genetic studies (see Ref. 28).

Jansen et al. (147) demonstrated that random deletion mutations in C. elegans could be generated in a large-scale manner using chemical mutagens. High-throughput methods for isolation of C. elegans deletion mutants have been described in detail (195). A C. elegans Knockout Consortium (http://elegans.bcgsc.bc.ca/knockout.shtml) consisting of three collaborating laboratories is currently working to produce strains possessing deletion mutations in all identified worm genes. Deletion strains are freely available to all investigators, and requests for knockout of a specific gene can be made online to the consortium (http://elegans.bcgsc.bc.ca/cgi-bin/submit_ko.pl).

E) RNA-MEDIATED GENE INTERFERENCE. One of the truly extraordinary experimental advantages of C. elegans is the relative ease by which gene function can be disrupted using dsRNA-mediated gene interference (RNAi). RNAi was discovered originally in C. elegans (90), but it now seems likely that the phenomena of posttranscriptional transgene silencing and quelling described earlier in plants and fungi, respectively, are occurring via dsRNA-regulated mechanisms (33, 253, 342). RNAi is thought to be a component of endogenous gene silencing mechanisms that protect cells against viral infection and transposon activity (22, 33, 142, 253, 293, 342). It may also play a regulatory role in developmental events (112, 171).

The precise molecular mechanisms that mediate RNAi are unknown. However, a number of important studies performed over the last 2 years have begun to provide a working model. dsRNA readily crosses cell membranes by unknown mechanisms. Once inside the cell, dsRNA is cleaved rapidly into small fragments ~21-25 bp in length. These small interfering RNAs (siRNAs) are the active agents and are thought to associate with a complex of proteins that includes RNases. The antisense strand on each 21-25mer base pairs with homologous single-stranded mRNA. After base pairing has occurred, the mRNA is destroyed by RNase activity (242, 355, 360). Recent studies have suggested that siRNAs may also serve as primers to transform target mRNA into dsRNA via the action of an RNA-dependent RNA polymerase (191, 234, 296). These newly synthesized dsRNAs are then cleaved into additional siRNAs. It has been estimated that just a few molecules of dsRNA in a cell are sufficient to dramatically knockdown targeted gene expression (90).

dsRNA disrupts targeted gene expression in a variety of invertebrate animals, plants, protozoans, and fungi (22, 33, 142, 253, 293). However, exposure of differentiated mammalian cells to long dsRNAs frequently triggers an interferon response leading to widespread, nonselective inhibition of gene expression and induction of apoptosis (156, 350). Recently, several groups have reported that the interferon response can be circumvented and specific gene silencing induced by transfecting mammalian cells with siRNAs (e.g., Refs. 43, 80, 122, 133, 236, 237, 244, 305, 354). It should be noted, however, that gene silencing by RNAi in mammalian cells is experimentally more complicated than in invertebrate cells. Mammalian cells must be transfected with siRNAs, and their silencing effect is relatively short-lived. In addition, the effectiveness of RNAi in silencing a mammalian gene may be dependent on the region of the mRNA that is targeted by the siRNAs (e.g., Ref. 133; see Ref. 211 for a review on the use of siRNA in mammalian cells).

The ability of dsRNA to readily cross C. elegans cell membranes is particularly advantageous for experimental disruption of gene expression. dsRNA can be microinjected anywhere in a worm and its effect will spread throughout the animal's body (90). In addition, RNAi can be induced by feeding worms bacteria making dsRNA (155, 320, 321) or by soaking them in dsRNA solutions (310). The systemic spread of gene silencing by dsRNA is dependent in part on a putative transmembrane protein SID-1 (351). Worms can also be engineered with inducible transgenes that are transcribed into dsRNA (317). In C. elegans cell culture, RNAi can be triggered simply by adding dsRNA to the culture medium (53; see also papers describing RNAi in cultured Drosophila cells, Refs. 42, 56, 262, 326).

RNAi is being used with great success to carry out large-scale, high-throughput reverse genetic screening of the C. elegans genome. Four separate groups have performed RNAi on ~33% of the animal's genes using microinjection (106, 250), feeding (94), and soaking (201) strategies. Of the ~6,275 genes examined to date, ~888 give rise to observable whole animal phenotypes when their expression is disrupted by dsRNA. As noted by Kim (163), the phenotypic analysis that has been carried out is at a "shallow level." Nevertheless, these important studies have assigned some level of function to a significant portion of the C. elegans genome and have provided the foundation for the generation and testing of countless new hypotheses.

F) DNA MICROARRAYS. DNA microarrays and DNA chips allow gene expression patterns to be monitored on a genome-wide scale. Affymetrix has produced DNA chips that collectively contain ~98% of the worm genome. Kim and co-workers (151, 269) have produced a microarray that contains ~94% of the worm's genes. The microarray was generated by PCR amplification of genomic DNA. Primer sequences are available online (http://cmgm.stanford.edu/~kimlab/primers.12-22-99.html), and the complete set of primers can be purchased from Research Genetics. In addition, investigators can send RNA samples to the Kim laboratory (http://cmgm.stanford.edu/~kimlab/wmdirectorybig.html) where they will be hybridized with the microarray and the data deposited online in the Stanford Microarray Database (http://genome-www5.stanford.edu/MicroArray/SMD/).

C. elegans microarrays and DNA chips have been used to examine developmental changes in gene expression (129, 151) and to identify oocyte- and sperm-enriched and sex-related genes (151, 269). Kim et al. (164) recently assembled data from multiple microarray experiments and developed an expression map of coregulated worm genes. Microarray technology in combination with cell culture and sorting methods (53) now make it feasible to identify cell-specific gene expression patterns under a variety of experimental conditions. For example, Chalfie and co-workers (361) have identified touch-cell specific genes by hybridizing a C. elegans microarray with amplified mRNA from FACS-enriched GFP-expressing touch neurons.

G) THE INTERACTOME: DEVELOPMENT OF A GENOME-WIDE PROTEIN-PROTEIN INTERACTION MAP. Protein-protein interactions play critical roles in most biological processes. Understanding these interactions is thus an essential component of functional genomics research. As in all organisms, protein-protein interactions can be identified in C. elegans by coimmunoprecipitation, pull-down assays, and yeast two-hybrid screens. Most physiologists are familiar with these experimental strategies and would typically use them when fishing for partners that interact with a single protein of interest. However, with the availability of complete genome sequences, large-scale, high-throughput yeast two-hybrid screens have been proposed as a way to map genome-wide protein-protein interactions (96, 183, 330). Genome-wide protein interaction maps can lead to new understanding of gene function in several ways. First, biological insights can be gained into proteins with unknown functions by linking them to proteins that operate in characterized biological pathways. Genome-wide screens can also identify novel interactions occurring between two or more proteins that function in a common physiological process, and they may identify new interactions and hence functions for previously characterized proteins.

Identification of genome-wide protein-protein interactions would clearly provide an extraordinary wealth of data from which experimentally testable hypotheses concerning the function and regulation of proteins, not only in C. elegans, but also in other organisms, could be proposed. Walhout et al. (335) (http://vidal.dfci.harvard.edu/main_pages/interactome.htm) recently assessed the feasibility of generating a protein-protein interaction map for the C. elegans genome. As a starting point, these investigators focused on 27 genes known to be involved in worm vulval development. PCR fragments were inserted into yeast two-hybrid vectors using the Gateway recombination cloning strategy (124, 336) described in section IIB2. Both the PCR and cloning steps are amenable to automation and thus can be scaled up to perform genome-wide interaction screens. Their screen of vulval development genes detected two novel interactions and ~50% of previously described interactions. When the vulval development genes were screened against a C. elegans cDNA library, 124 potential interactors were identified. Fifteen of these interactors had been identified previously in genetic screens, and 109 were new interactors predicted from genome sequence.

More recently, Davy et al. (66) began developing a protein-protein interaction map for the 26S proteasome from C. elegans. As a starting point, they searched the C. elegans genome for orthologs of human and yeast 26S proteasome subunits and related proteins. Thirty-two potential orthologs were identified, and 30 of these were PCR-amplified. Each of these ORFs was then screened against a C. elegans cDNA library using the high-throughput yeast two-hybrid assay described by Walhout et al. (336). Ninety-four potential interactors were identified. These included interactions reported for other organisms, novel interactions between proteasome subunits, and novel proteasome interacting proteins.

Boulton et al. (35) have combined large-scale protein-protein interaction mapping and RNAi feeding experiments to identify genes involved in the DNA damage response (DDR). BLAST searches were used initially to identify 75 putative C. elegans DDR orthologs. Sixty-seven of these orthologs were screened by high-throughput yeast two-hybrid assay, and 165 potential interactors were detected. The DDR orthologs and the interacting proteins were then subjected to RNAi feeding experiments (see sect. IIC3E). These studies confirmed that 12 of the 75 predicted orthologs were involved in the DDR response, and they identified 11 novel DDR genes. Taken together, the high-throughput screens carried out by Walhout et al. (335), Davy et al. (66), and Boulton et al. (35) have identified dozens of new genes and protein-protein interactions that provide the foundation for the development and testing of at least an equal number of new hypotheses.

H) ONLINE ACCESS TO C. ELEGANS BIOLOGY. The "worm community" is well known for its open sharing of data and reagents. As discussed above and summarized in Table 1, an extraordinary wealth of information on C. elegans is available publicly online. Indeed, the worm community was an early pioneer in the use of the internet for electronic data sharing.

WormBase (http://www.wormbase.org/) is a particularly noteworthy database. It provides an exhaustive catalog of worm biology. The database contains the worm genome, the complete cell lineages of hermaphrodite and male worms, genetic maps, gene expression patterns, RNAi phenotypes, worm genetics, and an atlas of worm anatomy.

I) VALUE OF GENOME-SCALE DATA COLLECTION. It is not uncommon to hear or read criticisms that are leveled at genome-wide investigative strategies such as microarray analyses, database mining, large-scale RNAi screens, and development of protein-protein interaction maps. For example, a recent editorial published in the Journal of General Physiology (7) stated that the "sheer amount of information" becoming available in the postgenomic era has forced a "shift toward high throughput screens, where the traditional notions of hypothesis-driven experiment and definitive mechanistic insight become elusive." Hypotheses are formed by asking questions about observations. One can make observations, for example, by measuring the kinetics of solute flux across a membrane or the electrophysiological properties of an ion channel, or by performing high-throughput, genome-scale screens. These data collection methods differ only in scale. The torrent of data made available by genome sequencing and genome level functional screens is unprecedented in biological research. Although these data may not be useful to all scientists, they are invaluable to anyone interested in integrative biology.

Rather than driving a shift away from hypothesis-driven investigation, genome-wide data and high-throughput screens now make it possible to develop and test hypotheses in ways unimaginable just a few years ago. Figure 6 illustrates this point schematically. Classical, hypothesis-driven investigations typically focus on one or a few proteins. Data on the function and regulation of these proteins are collected and hypotheses proposed to guide further experimentation and data collection. When these data are integrated with relevant genome-scale data, the results are inevitably the development of new hypotheses and deeper insights into biological function. For example, imagine that one is studying the role of binding protein X that regulates ion channel Y. A genome-scale protein interaction map might identify protein X as part of a cluster of interacting proteins (for example, see Refs. 35, 66). It would be reasonable to postulate that these interacting proteins are components of a signaling pathway and/or scaffolding complex that regulates channel Y. Similarly, microarray analyses might reveal that expression of binding protein X is coregulated with several other signaling/scaffolding proteins under specific physiological conditions, in different tissues, and/or at different times during development. Coregulation of these proteins would again suggest that they are components of a regulatory complex. Tissue-, physiological-, and developmental-dependent changes in binding protein X expression might suggest new functional roles for it as well as for channel Y. Physiologists interested in the molecular workings of complex systems from macromolecular protein assemblies to whole organisms should revel in data collected from genome-wide analyses.

View larger version (27K): [in this window] [in a new window]   Fig. 6. Integration of data from hypothesis-driven investigations and genome-wide analyses. The "classical" approach is envisioned as being primarily hypothesis driven and focused on the function and regulation of one or a few proteins involved in a biological process. Genome-wide analyses focus primarily on large-scale, high-throughput collection and collation of data. When relevant data from the two approaches are integrated, the result is inevitably deeper insight into biological processes and the development of new hypotheses.

	III. INTEGRATIVE PHYSIOLOGY OF CAENORHABDITIS ELEGANS MEMBRANE TRANSPORT PROCESSES

Top Previous Next References

In this section, I review the functional biology of membrane transport processes in C. elegans. It is important to stress that my discussion of the topic is not meant to be comprehensive. Instead, my intention here is to illustrate the experimental power of genomically defined and genetically tractable nonmammalian model organisms such as C. elegans for the study of integrative membrane transport physiology. I do this by describing how the molecular workings of complex physiological processes in which ion channels and transporters play key roles have been and are continuing to be elucidated in C. elegans through a combination of forward and reverse genetic screening methods and direct physiological, behavioral, and cell biological measurements. Where appropriate, I discuss how studies on nematodes have provided new insights into ion channel and transporter physiology in mammals.

A.  Worms Have Feelings Too: DEG/ENaC Cation Channels

1.  Mechanosensation in C. elegans

All organisms have the ability to detect and respond to mechanical force. Mechanical signaling plays an important role in a wide variety of biological processes ranging from osmoregulation to hearing and locomotion (105, 118). Remarkably, relatively little is known about the molecular basis of mechanosensation.

C. elegans has proven to be an invaluable model system in which to identify genes encoding proteins required for mechanosensory processes. Mechanical signals likely regulate a variety of nematode behaviors including locomotion, foraging, and the ability of the animal to detect and respond to touch. The best understood mechanosensory behavior is the avoidance response initiated when the animal is touched lightly on the body with an eyelash hair (45). Gentle touch to either the posterior or anterior body causes the animal to move away in a forward or backward motion.

Laser ablation studies identified five neurons required for the response to gentle body touch (46). These "touch" neurons are also called microtubule cells because their processes contain a bundle of microtubules, each of which is formed from 15 protofilaments (47, 48). Touch neurons include the anterior ventral microtubule cell (AVM) and the anterior and posterior lateral microtubule cells left, right (ALML/R and PLML/R, respectively). The posterior ventral microtubule cell (PVM) is also considered to be a touch neuron because its morphology is the same as that of the AVM, ALML/R, and PLML/R and because it expresses touch neuron specific genes. However, ablation of this neuron does not appear to alter touch sensitivity (46).

Genes required for touch neuron mechanosensory functions were identified by mutagenizing worms and assessing their ability to respond to gentle body touch (44, 45). Touch-insensitive mutant worms are mechanosensory abnormal, and the genes responsible for this abnormality are termed mec. Approximately 15 mec genes have been identified to date. Several of these genes are required for normal touch neuron development. At least eight mec genes encode proteins that have been postulated to form a mechanosensitive ion channel complex (203, 314, 315).

The mec-4 and mec-10 genes encode ion channel forming proteins that share significant homology with epithelial Na⁺ channels or ENaCs (5, 203). The first ENaC-like channel encoding gene identified in C. elegans was deg-1 (degeneration of certain neurons). Gain-of-function mutations in deg-1 cause various nematode neurons to swell and degenerate (50); hence, the protein encoded by the gene was termed a "degenerin." One particular mutation in deg-1 results in the substitution of alanine at position 707 to valine (102). This amino acid is located at a putative extracellular site adjacent to the second transmembrane domain. Mutation of the equivalent alanine in mec-4 and mec-10 also induces neuronal degeneration (72, 138). Heterologous expression studies of Drosophila and vertebrate DEG/ENaC homologs have demonstrated that this mutation induces constitutive channel activation (1, 2, 51, 334).

Twenty-five members of the DEG/ENaC family have been identified in the C. elegans genome (315). The channels are expressed in a wide variety of cell types including neurons, muscle cells, intestine, and hypodermis (313, 314). Three of these channels, UNC-105, UNC-8, and DEL-1, have been suggested to play roles in processes thought to be regulated by mechanical force (101, 193, 241, 294, 316).

The mec-7 (117, 288) and mec-12 (97) genes encode - and -tubulins, respectively. Mutations in these genes disrupt the formation of the 15 protofilament microtubules in touch neurons (48, 97, 289). The mec-2 gene encodes a stomatin-like protein (139). Mammalian stomatin is a 31-kDa integral membrane protein that associates with the cytoskeleton (298) and lipid raft proteins (285). Red blood cells in patients with hereditary stomatocytosis are deficient in stomatin and exhibit abnormalities in shape and monovalent cation permeability (68). Interestingly, stomatin is coexpressed with ENaC-type channels in mammalian mechanosensory neurons (95, 103, 202).

Touch neuron processes are surrounded by an extracellular matrix or "mantle," which attaches the cells to the cuticle (45). Mantle proteins play important roles in touch neuron mechanosensation. The mec-1 gene encodes a protein required for formation of the mantle (45), and mec-5 encodes a specialized collagen that is secreted by hypodermal cells (74). Genetic analyses suggest that MEC-5 may interact with MEC-4 and MEC-10 (113). The mec-9 gene encodes a protein secreted by touch neurons that interacts genetically with MEC-5 (74). MEC-9 contains six epidermal growth factor-like repeats, five Kunitz-type protease inhibitor domains, and a glutamic acid-rich region, suggesting that it may play a role in mediating protein-protein interactions.

Goodman et al. (108) demonstrated recently that MEC-4 and MEC-10 form functional Na⁺ channels and that MEC-2 plays an important role in regulating their activity. When expressed alone or together in Xenopus oocytes, wild-type MEC-4 and MEC-10 do not give rise to detectable Na⁺ currents. However, amiloride-sensitive currents are detected when channel proteins containing the "degenerin" mutation, MEC-4d and MEC-10d, are coexpressed. Coexpression of MEC-2 with MEC-4d and MEC-10d increases current ~40-fold. Currents are also detected, albeit at a much lower level, when MEC-2 is expressed with wild-type MEC-4 and MEC-10.

Genetic studies suggest that MEC-2 interacts with the MEC-4/MEC-10 channel complex (113, 138). In support of this finding, Goodman et al. (108) demonstrated that anti-MEC-2 antibodies immunoprecipitated MEC-4d and MEC10d expressed in oocytes. Taken together, these and other studies indicate that MEC-4 and MEC-10 form a heteromeric ion channel that is regulated by MEC-2 (Fig. 7).

View larger version (32K): [in this window] [in a new window]   Fig. 7. Model of a mechanosensory ion channel complex in C. elegans touch neurons. Mechanical distortion of the cuticle is postulated to gate the MEC-4/MEC-10 channel complex open by displacing it with respect to cytoskeletal and extracellular matrix attachment points.

The central domain of MEC-2 (amino acids 114-363) is 64% identical to human stomatin. Of 54 mutant alleles of mec-2 detected in genetic screens, more than half map to this region, indicating that it is particularly important for mechanosensory neuron function (139). Consistent with this idea, the central domain alone is capable of stimulating amiloride-sensitive currents, albeit significantly less than the full-length protein. Human stomatin induces a similar increase in amiloride-sensitive currents. Taken together, these studies provide the first direct evidence that stomatins regulate ion channel activity. MEC-2 has no effect on membrane trafficking of MEC-4d or MEC-10d, suggesting that it functions to increase channel open pr