(Note: this text does not include figures)

· An overview of cystic fibrosis

 It has been known for more than 50 years that cystic fibrosis (CF) is a genetic defect that leads to the accumulation of mucus in many exocrine and exocrine-associated organs, but it is only in the last decade that the basis of CF was traced to a defect in a gene that codes for a chloride channel called CFTR, (cystic fibrosis transmembrane conductance regulator). Although CFTR is known to be a chloride channel, the way that CFTR works and the consequences of its absence or malfunction are presently under intense investigation. CFTR is found primarily in wet epithelia, consistent with the symptoms that define CF. Therefore, a study of epithelia, and especially CFTR's contributions to epithelia function, is a prerequisite to understanding CF disease.

· Extent and functions of epithelia 

From the first milk we drink to the last tears we shed, we rely on epithelial transport. By sheer volume alone, transporting epithelia (wet epithelia) announce their importance. The body surface is less than 2 m2. By comparison, the wet, mucosal, transporting epithelial tissues that line our digestive, respiratory, renal, reproductive and sweat gland systems are immense. The alveolar surface of the lungs alone is 70 m2; the surface area of the small intestine is 180 m2; the total length of the kidney tubules is 80,000 m; and our 2.3 million sweat glands comprise 5,300 m of secretory tubules and about the same length of reabsorbtive duct. Although hidden, these vast surfaces and miles of tubules are essential for both maintaining and propagating life, and they loom large in medical physiology. All transporting epithelia are multiply controlled by hormonal and direct neural innervation.

The surface area of our internal organs is surprisingly large because numerous wrinkles have been used to increase the surface area relative to the volume. These wrinkles range from simple ridges to fine, hair-like extensions called villi or invaginations of various complexity called glands. But these structures shouldn't mislead us from the basics: epithelia are ultimately sheets of cells only one cell thick.

A typical mucosa or fluid transporting cell layer is a monolayer of cells, each one welded by tight junctions around its entire circumference to all of its immediate neighbors. Tight junctions make net transport of ions and fluids across epithelial sheets possible by forming a barrier that limits the diffusion of ions around the outside of the cells (Fig. 1). The pathway around the cells and through the tight junctions is called the paracellular pathway. Organs differ greatly in the barrier functions of the tight junctions--some organs (like the gall bladder) allow ions to flow easily through the pathway around the cells, other epithelia (like the sweat duct) are very tight to the flow of ions around the cell.

Tight junctions also separate the cell membrane into two distinct types: an apical membrane that faces the outside of the body, and a basolateral membrane that faces the blood. These membranes have different lipid compositions and, most importantly, different populations of membrane proteins. This polarization of epithelial cells sets them apart from most other cells in the body, and enables them to transport substances in a directed fashion across the cell sheet.

Epithelial sheets usually work as a unit. In fact, the general rule that cells are separate individuals is modified somewhat in epithelia because the cytosol of adjacent cells is linked by gap junctions that permit electrical current and small molecules to diffuse freely among the cells.

Some epithelial are made up of multiple copies of essentially identical cells, while other epithelia are made of many different kinds of cells that perform specialized functions. For example, some epithelial cells are ciliated and use much of their cellular energy to generate mechanical movements of the cilia, other cells are specialized to secrete mucus, while still others are primarily concerned with electrolyte transport.

· A model system to study electrolyte transport

 To understand fluid secretion in overview, consider the model epithelial system formed by T84 cells, a line of human colonic tumor cells. These cells are very similar to the fluid-secreting epithelium found in the crypts of the intestines. The basic transport scheme is shown in the figure. At rest, the cell has a fairly low conductance to ions (channels are mostly closed).

Fluid secretion in the intestine and in T84 cells is induced by agents that increase cAMP, cGMP, or Ca2+ . In T84 cells, maximal fluid secretion is produced by opening Cl- channels in apical membrane via cAMP-dependent protein kinase (PKA) and opening Ca2+-activated K+ channels in the basolateral membranes. As shown, each of those ions then leaves the cell on opposite sides, (that is, there is a net loss of KCl, but it is vectorial). The negatively charge of Cl- causes Na+ to follow via the paracellular pathway, resulting in a net secretion of NaCl; water follows to maintain osmotic balance. The passive loss of KCl is replenished by the cooperative actions of Na+, K+-ATPase and Na+, K+, 2 Cl- cotransporter.

Secretory diarrhea, estimated to cause 5-10 million deaths annually in Africa, Asia and Latin America, is caused by infection with a variety of bacteria or protozoa, notably E. coli, Salmonella, Vibrio cholerae, and Giardia. Cholera toxin is known to exploit features of the signal transduction pathway in crypt cells to induce massive, continuous secretion. Cystic fibrosis, in contrast, eliminates secretion via these cells.

· Electrolyte transport: basic components

 Epithelial cells, like all cells, are surrounded by a plasma membrane made up primarily of phospholipids, with embedded and adsorbed proteins and glycolipids. Pure lipid membranes are modestly permeable to water, but are highly impermeable to charged molecules (ions). The only way that ions such as sodium (Na+) potassium (K+) and chloride (Cl-) can cross a membrane is via special molecules that either ferry them across the membrane by shielding their charge, or that provide a tiny, water-filled pore through the membrane.

A host of complex proteins has evolved to move ions across membranes. The study of these proteins is one of the most vigorous areas of modern biology. Four general kinds of ion transport molecules work together in all cells to effect ion transport:

· Pumps or transport ATPases use the energy of ATP hydrolysis to move ions against a concentration gradient. These are the fanciest of the transport molecules in terms of what they accomplish: they link, at the level of a single molecule, energy consumption with the vectorial movement of a substance. The premier example of such a molecule is the sodium pump or Na+, K+-ATPase, which is the primary source of energy for almost all ion transport.

· Exchangers alternately bind ions on opposite sides of the membrane. In this way they couple the flow of an ion one way across the cell membrane with another ion flowing in the opposite direction. The best known examples of exchangers are the anion exchanger AE1, or Band 3, that couples bicarbonate and chloride exchange in red blood cells, and NHE or the Na+/H+ exchanger that is a key component of pH regulation in cells.

· Cotransporters, especially Na+-coupled cotransporters, simultaneously bind two or more ions and then move them together across the membrane. For example the Na+/K+/2 Cl- cotransporter (there is no simple name for this one!) is responsible for pushing Cl- uphill into the cell. In polarized epithelia, the transporter is usually located on the basolateral membrane of secreting epithelia and on the apical membrane of reabsorptive epithelia. Furosemide inhibits, bumetanide inhibits 50-100 times better. These `loop' diuretics are negatively charged at physiological pH.

· Channels form water-filled pores through the membrane, with binding sites for ions. The fundamental difference between ion channels and other transporters is the rate at which they allow ions to traverse the membrane. An ATPase might move 100-300 ions per second, while a very rapid exchanger might move 40,000-100,000. However, even a low conductance channel can conduct more than a million ions per second, and because of this the ionic current of many single ion channels can be detected with special methods (patch-clamping). A wealth of information has resulted from this.

· Ion channels in more detail

 Ion channels are described by their ion selectivity, type of gating, conductance, gating kinetics, and the types of molecules that block them.

Ion selectivity is the most fundamental characteristic of a channel, and most channels are named accordingly, e.g. sodium channels and chloride channels. Gating is also a fundamental property of a channel, and the gating mechanism is often added as a modifier of the name, e.g. calcium activated potassium channel or voltage gated sodium channel. Three major types of gating are voltage gating, ligand gating, and gating by intracellular messenger.

Channel kinetics have only been accessible since the advent of patch clamp recording. A curious feature of channels is that they are typically either completely open or completely closed, and they switch almost instantaneously between these rates with patterns that are on average, characteristic for each type of channel.

Unit conductance is a property that has only been emphasized since the advent of single channel recording (see below). Channels can have remarkably high conductances. The theoretical upper limit for a channel in physiological solutions is about 300 pS (Hille, 1984); this value is almost reached by several real channels.

Blocking agents have been discovered for many channels, and are extremely important tools in the analysis of channel function. Some channels, in fact, are first differentiated from one another by their relative susceptibilities to specific blocking agents. For example, a well known channel blocker among epithelial biologists is amiloride, which blocks certain sodium channels with high affinity.

Comparison of carriers and channels. As noted above, a great disparity exists in terms of the small ionic currents that can be generated by single ATPase molecules vs. the large currents generated by single ion channels. This disparity is reciprocally reflected in the relative densities of channels and ATPases. ATPases in exocrine glands can reach densities of 8,000 per m m2, whereas estimates of the total number of K+ channels in similar cells can be as low as 25! In other calculations of this sort, acinar cells in the guinea pig pancreas bind about 780,000 ouabain molecules per cell (one per sodium pump); ductal cells bind about 2,000,000. The very small number of ion channels that can have large physiological effects has bedeviled attempts to visualize ion channels with antibodies and to isolate channel proteins with biochemical methods.

· Transport protein structure

 The primary amino acid sequences of transport proteins as determined by gene cloning is transforming the study of ion transport from a branch of physiology to a branch of molecular biology. In 1985, sequences were known for just two channels and a few transporters. Now, the sequences of all of the major transport proteins involved in epithelial transport are known, and new transporters are being found at a rapid rate.

For example the structures of three of the four major transport proteins involved in Cl- secretion by T84 cells are known.

· Na+, K+-ATPase consists of an a and b subunits. The a subunit has 1,016 amino acid residues and appears to traverse the membrane 8 times; the b subunit has 302-304 amino acids, is heavily glycosylated, and traverses the membrane once.

· The Na+/K+/2 Cl- cotransporter is a 1,191 amino acid polypeptide with 12 putative membrane spanning domains and 9 potential sites for N-linked glycosylation; the protein is heavily glycosylated (core protein c. 135 kDa; glycosylated protein c. 165 kDa. The protein has very little homology with any other know protein except the Na/Cl cotransporter.

· The cystic fibrosis transmembrane conductance regulator (CFTR) was cloned in 1989, and since then a great deal has been learned about its structure and function.

· CFTR 

The gene for CFTR consists of almost a quarter of a million base pairs, which generate, after splicing, an mRNA transcript of ~6500 base pairs that codes for a 1480 amino acid integral membrane protein. The structure of CFTR consists of two sets of 6 transmembrane spanning regions, each associated with an nucleotide-binding domain, and separated from each by a large, highly charged region containing numerous consensus sequence sites for phosphorylation by PKA, protein kinase C (PKC) and other kinases. The first set of 6 transmembrane-spanning regions, which are probably mostly alpha-helices, forms part of the pore region of the channel. When the Cl- ion conducting pore of CFTR is open, Cl- ions pass through the channel at a high rate. What controls the gate that guards the pore?

Models of CFTR structure and function. The model on the left shows the original cartoon of a CFTR channel, with charged amino acids indicated by + and - signs, PKA phosphorylation sites indicated by s and PKC sites indicated by Ñ . On the right is a model of CFTR regulation by phosphorylation and nucleotides. from Riordan et al., Science, 245, 1066, 1989; and from Hwang et al., P.N.A.S., 91, 4698, 1994.

The complex structure of CFTR was initially bewildering, because the nearest relatives of CFTR are active pumps that use the energy of ATP hydrolysis to move substances against a concentration gradient. Yet all evidence pointed to CFTR being a passively conducting ion channel. The mystery has been partially solved by a series of elegant experiments which, in aggregate, indicate that CFTR uses ATP hydrolysis to open the channel (a switch or gate mechanism) and then as a timer mechanism to determine how long the channel stays open before closing. Added to this complexity is the requirement for phosphorylation, which can be considered to be another gate that protects CFTR from being opened inadvertently.

A working hypothesis for CFTR function is shown in the figure. In this model, the probability of ATP binding is increased by phosphorylation of the R domain. When ATP binds to one NBF, it is hydrolyzed and the channel opens briefly. The channel might bind another ATP and reopen again, or might be dephosphorylated and close for a longer period. If it is sufficiently phosphorylated, ATP can be bound by the second NBF, which stabilizes the open state. In some conditions, CFTR can remain continuously open for tens of seconds or minutes, either because the ATP hydrolysis rate is somehow slowed, or because the NBF is occupied by a poorly hydrolyzed substrate. This condition of very long open times is, in fact characteristic of CFTR is several epithelia--in these cells CFTR appears to be constitutively active.

This model assumes that CFTR channels are resident in the membrane. However, in some epithelia transport is also regulated by moving transporters into and out of the membrane via exocytosis/endocytosis. The role that such a process plays in CFTR-mediated transport is presently under investigation.

· Mutant CFTR

In the 5 years since the CF gene was cloned, more than 400 putative mutations have been reported to cause cystic fibrosis. About half of these mutations either introduce a stop codon or a frame shift or alter splicing, so that essentially no functional protein is produced, while the other half are missense or (rarely) in-frame deletions that alter the amino acid sequence of the protein. The large proportion of missense mutations that lead to disease may be partly explained by the unusual sensitivity of the protein to conditions that affect folding or maturation. Even unmutated CFTR is processed with low efficiency, and missense mutations often result in an almost total absence of CFTR protein from the membrane. The most prevalent mutation giving rise to CF, D F508, causes misprocessing of CFTR. If mutated protein is processed to the membrane, it may fail to function properly because it fails to bind or hydrolyze ATP properly, resulting in a channel that is almost always closed. A mutation of this type is G551D. Such mutations may cause only partial loss of function of CFTR, such as G551S. Finally, some mutations appear to interfere with the flow of Cl- ions through the channel, resulting in lower conductance channels. Examples are R117H, R334W, and R352Q. These conductance mutants are in some ways the most informative, because they establish definitively that the CF syndrome is related to altered Cl- conductance. They also provide some evidence about the location of the pore. It is interesting that conductance mutations typically give rise to milder forms of cystic fibrosis.

  · Summary

 Cystic fibrosis is caused by any mutation that severely interferes with the functioning of CFTR, a Cl- channel that is regulated by cytosolic nucleotides and by phosphorylation. The loss of Cl- conductance in many wet epithelia has widespread consequences, many of which are still poorly understood. The possibility that all CF pathology arises from loss of Cl- conductance, vs. the possibility that CFTR has additional, non-channel functions is considered in a later session.

 · Selected Readings

Combined readings for this section and the section on Organ Specific Defects in Cystic Fibrosis are found at the end of the latter section.