Glucose Transport in the Intestine, Kidney, Fat, and Muscle

A chemiosmotic cycle transports glucose uphill from the lumen of the intestine to the blood (Fig. 11-2). Tight junctions restrict movement of glucose between the epithelial cells, so all of the glucose must move through the cytoplasm. Glucose transport across the epithelial cells requires the following components:

The molecular composition of the membrane domains explains the mechanism of glucose transport. Na⁺K⁺-ATPases use ATP hydrolysis to produce Na⁺ and K⁺ gradients across the plasma membrane by continuously pumping Na⁺ out of and K⁺ into the cell. SGLT Na⁺/glucose symporters use Na⁺ moving inward down its electrochemical gradient to accumulate high internal concentrations of glucose from the lumen. In this step, energy is expended (dissipation of the Na⁺ gradient) to move glucose uphill. GLUT uniporters in the basolateral membrane simply facilitate movement of cytoplasmic glucose down its concentration gradient out of the cell. In the gut, this process provides for the uptake of glucose from food. Renal proximal tubule cells use a similar strategy to recapture glucose filtered from blood, transporting it across the tubule cell and back into the blood.

Glucose uptake by fat and muscle cells offers a different perspective. These tissues are designed to take up glucose from the blood when it is plentiful following a meal. Mammals have genes for six isoforms of the classical D-glucose uniporter. Insulin in the blood regulates the availability of the GLUT4 isoform in the plasma membrane. Muscle and fat express GLUT4 but store it internally in membrane vesicles. After a meal, high blood glucose stimulates secretion of insulin into blood. Signal transduction mechanisms (see Fig. 27-7) lead to fusion of these GLUT4 vesicles with the plasma membrane. That increases the rate of glucose transport into fat and muscle by 5-fold to 20-fold, lowering the blood glucose concentration and providing these cells with glucose, which they then convert to glycogen and triglycerides for storage.

Salt and Water Transport in the Kidney

In a section of the kidney tubule called the loop of Henle, the epithelium uses Na⁺K⁺-ATPase pumps and Na⁺/K⁺/2Cl⁻ symporters to reabsorb NaCl that is filtered from blood into the excretory pathway (Fig. 11-3). Without this provision, salt would be lost in urine. Tight junctions seal this epithelium, so that salt must pass through the cells to return to the blood. Na⁺/K⁺/2Cl⁻ symporters in the apical plasma membrane provide a way for NaCl to enter the cell down its concentration gradient. Abundant Na⁺K⁺-ATPases in the basolateral plasma membrane (5000/mm²) create a Na⁺ gradient to drive the symporter and to clear the cytoplasm of Na⁺ accumulated from the tubule lumen. KCl that enters with Na⁺ through the Na⁺/K⁺/2Cl⁻ symporter leaves the cell through channels: K⁺ channels in apical and basolateral membranes and Cl⁻ channels in basolateral membranes.

Figure 11-3 sodium chloride transport by the epithelium of the kidney tubule. Tight junctions seal the space between these polarized epithelial cells of the thick ascending limb of the loop of Henle. Na⁺K⁺-ATPase pumps (space-filling model) in the basolateral plasma membrane drive Na⁺/K⁺/2Cl⁻ symporters in the apical plasma membrane. K⁺ channels in the apical plasma membrane and K⁺ channels and Cl⁻ channels in the basolateral plasma membrane provide paths for K⁺ to circulate and for Cl⁻ to follow Na⁺ across the cell from the lumen of the tubule to the blood compartment.

A drug that is used to treat congestive heart failure—furosemide—inhibits the Na⁺/K⁺/2Cl⁻ symporter in the loop of Henle. A weak heart leads to accumulation of fluid in the lungs (causing shortness of breath) and other tissues (causing swelling of the ankles). Inhibition of the Na⁺/K⁺/2Cl⁻ symporter reduces NaCl reabsorption, so the kidney produces large quantities of urine, clearing excess fluid from the body and relieving symptoms.

Cystic Fibrosis as a Transporter Disease

Normally, cells in the lung and gastrointestinal tract use a complicated selection of familiar pumps and carriers to secrete salt and water at their apical sur-faces (Fig. 11-4). Na⁺K⁺-ATPases in the basolateral membrane set up an electrochemical gradient of Na⁺, which is exploited by basolateral membrane Na⁺/K⁺/2Cl⁻ symporters to take in Na⁺, along with K⁺ and Cl⁻ anions. The inward movement of Na⁺ down its electrochemical gradient drives the entry of K⁺ and Cl⁻ up their gradients. This brings excess potassium chloride into the cell. (The K⁺ brought in by both the Na⁺K⁺-ATPase and the Na⁺/K⁺/2Cl⁻ symporter recycles after exiting from the cell by way of channels in the basolateral plasma membrane. Thus, K⁺ is merely catalytic.) Excess Cl⁻ is left inside the cell. The cystic fibrosis transmembrane regulator (CFTR) protein, an ABC pump, in the apical plasma membrane acts as a Cl⁻ channel. When protein kinase phosphorylates the regulatory domain and ATP binds to the cytoplasmic domains, a conformational change opens a Cl⁻ channel across the membrane. Cl⁻ moves down its electrochemical gradient out of the cell, carrying charge to the outside. The whole epithelium becomes polarized, with the lumen electrically negative relative to the extracellular fluid compartment. This electrical driving force allows Na⁺ to move between cells, from the extracellular fluid compartment through leaky tight junctions to the surface of the epithelium. Sodium chloride on the apical surface creates an osmotic force that draws water down its concentration gradient across the cells to the outside through water channels (see Fig. 10-15). CFTR also appears to inhibit the transport mechanisms that reabsorb fluid from the lumen of the epithelium. A balance between this fluid secretion and fluid reabsorption normally keeps the surface of the epithelium properly hydrated, allowing the cilia to clear the lung of bacteria and secretions and the ducts of the pancreas to secrete digestive enzymes.

Figure 11-4 salt and water transport across the epithelium lining the respiratory tract. Leaky tight junctions partially seal the space between these polarized epithelial cells. Na⁺K⁺-ATPase pumps in the basolateral plasma membrane drive Na⁺/K⁺/2Cl⁻ symporters in the basolateral plasma membrane. CFTR Cl⁻ channels in the apical plasma membrane allow Cl⁻ to move into the lumen, creating a negative electrical potential that pulls Na⁺ between the cells into the lumen. CFTR also releases ATP, which activates additional Cl⁻ channels. Water follows sodium chloride into the lumen through water channels and between the cells. Basolateral K⁺ channels allow K⁺ to circulate.

Patients with cystic fibrosis have mutations in CFTR that result in defects in apical Cl⁻ transport and secretion. Their lungs are too dry as a result of this imbalance in fluid secretion and reabsorption on the surface of epithelia. This situation is life-threatening because cilia in the respiratory tract cannot move sticky, dry, mucus containing bacteria and viruses out of the lungs, thereby predisposing to respiratory infections. Sticky secretions in the pancreatic ducts also interfere with the secretion of digestive enzymes by the pancreas. Many different mutations in the CFTR gene cause the disease. The most common mutation (67% of cases) deletes the codon for phenylalanine 508 (F508). The resulting protein is temperature-sensitive, not folding properly at 37°C and failing to negotiate the secretory pathway to the plasma membrane. Patients with two copies of this mutation on chromosome 7 have classic cystic fibrosis. Heterozygotes with one normal gene (about 5% of the population) have no symptoms. Patients who have a combination of this DF508 mutation with a number of other mutations in the other copy of the gene vary in the severity of their pancreatic problems but still have typical lung disease.

Description of an Action Potential

If a microelectrode (see Fig. 10-16A) drives a small positive or negative current into a cell, a second microelectrode a short distance away detects a small voltage response. These electrotonic potentials decline rapidly with distance if the cell in nonexcitable.

In striking contrast to these small local currents, when the plasma membrane of an excitable cell, such as neuron or muscle, is depolarized beyond a certain level, called a threshold, the membrane responds over a few milliseconds with a large, stereotyped change in membrane potential, called an action potential (Fig. 11-6). Voltage-gated ion channels (see Fig. 10-7) generate this powerful electrical signal that spreads rapidly (10 m/sec) over the entire plasma membrane. During an action potential, the membrane potential can reach a peak of +40 to 50 mV before repolarizing to the resting potential. Because action potentials are self-triggering, they travel without dissipation over long distances. This high-speed transmission is very efficient, requiring movement of very few ions across the membrane.

Figure 11-6 the time course of an action potential passing a measuring electrode inserted through the plasma membrane of a squid giant axon. Spread of an action potential from an adjacent area of the membrane brings the membrane potential E_m, to threshold, triggering the action potential at this point on the membrane. The other curves show the conductance of the membrane at this point for Na⁺ and K⁺ expressed as the concentration of open channels. The lower trace shows the times during which K⁺ leak channels open and close at this point. E_Na is the Na⁺ equilibrium potential, and E_K is the K⁺ equilibrium potential.

The molecular events during an action potential were first characterized around 1950 in squid giant axons using microelectrodes coupled to an electronic feedback circuit. This clever “voltage clamp” holds the membrane potential constant by providing the cell with electrical current to compensate for changes in ion currents. Investigators discovered that changes in permeability to Na⁺ and K⁺ ions produced action potentials. Changing one variable at a time, they determined the time and voltage dependence of ion-specific conductance. They also determined the relationship between conductance and voltage. From these relationships, measured under controlled conditions, they could calculate the membrane response to virtually any experimental condition. To explain these changes in permeability, they postulated the existence of ion channels. The voltage clamp provided a direct measure of this channel activity. This approach also revealed the behavior of channels held at a potential more positive than their resting potential.

Three Channels Generating Action Potentials

Voltage-gated Na⁺ and K⁺ channels open and close in sequence to produce action potentials. Depending on the type of open channel, the membrane potential varies in time between the K⁺ equilibrium potential (E_K) and the Na⁺ equilibrium potential (E_Na) (see Fig. 10-18). Because membrane depolarization activates these ion channels, and because the response spreads this depolarization, triggering an action potential initiates a cascade of reactions that moves over the membrane, first to depolarize and then to repolarize the membrane. In nerves, just three types of voltage-gated channels are required to generate action potentials:

The properties of these channels explain the time course of an action potential as follows:

Stage 1: At rest, the membrane is slightly permeable to K⁺ but not to other ions, so the resting potential is near E_K, about -70 mV. K⁺-selective leak channels and a few open voltage-gated K⁺ channels contribute to this basal K⁺ permeability.

Stage 2: If the membrane is depolarized by an oncoming action potential and reaches the threshold potential, K⁺-selective leak channels close and voltage-gated Na⁺ channels open. Because the membrane is permeable only to Na⁺ and because many Na⁺ channels open, Na⁺ moves into the cell and the membrane potential rapidly approaches E_Na, about +45 mV.

Stage 3: After 1 to 2 msec, Na⁺ channels spontaneously inactivate and slowly responding delayed-rectifier K⁺ channels open. Now the membrane is strongly and selectively permeable to K⁺, so K⁺ moves out of the cell, and the membrane potential reverses all the way to E_K, about -80 mV. K⁺ channels are less synchronized than Na⁺ channels, so the membrane potential falls more slowly than it rises.

Stage 4: Delayed-rectifier K⁺ channels close progressively as the membrane repolarizes, and K⁺-selective leak channels open, returning the membrane potential to the resting voltage, just above E_K.

During an action potential, the membrane voltage changes by 100 to 150 mV in 1 to 2 msec. The membrane bilayer is approximately 7 nm thick, so this voltage corresponds to a field variation on the order of 150,000 volts/cm in 1 to 2 msec. Such strong forces elicit conformational changes in membrane proteins, such as voltage-gated ion channels.

Membrane Depolarization: The Stimulus for Action Potentials

The initial depolarization of the plasma membrane that triggers an action potential can arise from activation by a neurotransmitter (see the section that follows) or spread of an action potential from an adjacent membrane or from an adjacent cell through a gap junction. Membrane depolarization must exceed a certain threshold to trigger an action potential. The threshold arises directly from the properties of the ion channels. Depolarization less than threshold activates a few Na⁺ channels, producing a small inward Na⁺ current, but it also activates some delayed-rectifier K⁺ channels, resulting in K⁺ efflux. If the Na⁺ conductance is small in relation to the K⁺ conductance, outward currents predominate and the membrane repolarizes. Depolarization greater than threshold activates additional Na⁺ channels, yielding inward Na⁺ currents greater than outward K⁺ currents, at least briefly. This positive feedback loop further depolarizes the membrane, amplifying activation of Na⁺ channels and producing the cascade of channel activation that makes action potentials an all-or-nothing event.

Neuromuscular Junction

Motor neurons in the spinal cord and brainstem control contraction of skeletal muscle cells (see Fig. 39-14). Long axons from these neurons terminate in synapses on skeletal muscle cells, called neuromuscular junctions (Fig. 11-8A–B). Every neuronal action potential that reaches a neuromuscular junction evokes an action potential that spreads over the postsynaptic surface of the muscle cell and initiates contraction. This highly reliable, one-to-one communication depends on chemical transmission by acetylcholine between the nerve and muscle. Highly concentrated nicotinic acetylcholine receptors in the postsynaptic membrane (˜20,000/mm²) transduce the arrival of extracellular acetylcholine into membrane depolarization.

Figure 11-8C illustrates the membrane proteins required for neuromuscular transmission. Both the nerve terminal and muscle depend on Na⁺K⁺-ATPase and Ca²⁺-ATPase pumps to maintain gradients of Na⁺, K⁺, and Ca²⁺ across their plasma membranes. Both presynaptic and postsynaptic cells need voltage-gated Na⁺ channels and K⁺ channels for action potentials. Additionally, the presynaptic membrane requires voltage-gated Ca²⁺ channels to trigger secretion of acetylcholine.

A neuronal action potential initiates synaptic transmission by admitting Ca²⁺ into the presynaptic terminal through voltage-gated Ca²⁺ channels. Within less than 1 msec, Ca²⁺ triggers fusion of synaptic vesicles containing acetylcholine with the plasma membrane. Within microseconds, acetylcholine released into the synaptic cleft between the cells reaches millimolar concentrations and binds postsynaptic acetylcholine receptors.

Weak but cooperative binding of acetylcholine to two subunits of the acetylcholine receptor (see Fig. 10-12) opens a nonselective cation channel. The open pore is about equally permeable to K⁺ and Na⁺ and less permeable to Ca²⁺, so the membrane potential collapses toward a reversal potential (see the section titled “Consequence of Multiple Channel Types Opening Simultaneously”) of about 0 mV. This is above threshold for triggering a self-propagating action potential in the muscle plasma membrane, which occurs with nearly 100% efficiency. The action potential traveling over the muscle plasma membrane activates voltage-sensitive Ca²⁺ channels that trigger Ca²⁺ release from smooth endoplasmic reticulum, resulting in contraction (see Fig. 39-15).

Two different mechanisms terminate activation of acetylcholine receptors. First, an extracellular enzyme, acetylcholinesterase, rapidly degrades free acetylcholine, usually depleting acetylcholine from the synaptic cleft in a few milliseconds. Acetylcholine then dissociates from the receptor, closing the channel. Second, during prolonged exposure to acetylcholine, a conformational change in the acetylcholine receptor increases its affinity for bound acetylcholine and closes the channel. In this desensitized state, acetylcholine dissociates only when its extracellular concentration is very low. Once acetylcholine dissociates, the receptor slowly returns to rest.

Nerve terminals retrieve synaptic vesicle membrane by endocytosis (see Chapter 22). Cytoplasmic enzymes synthesize new acetylcholine. A V-type ATPase proton pump (see Fig. 8-5C) acidifies the lumen of synaptic vesicles, providing an electrochemical potential to drive an acetylcholine/H⁺ antiporter, which concentrates acetylcholine in vesicles.

Central Nervous System Synapses

Synaptic transmission between neurons in the CNS (Fig. 11-9) differs fundamentally from the efficient, one-to-one coupling at neuromuscular junctions, where every presynaptic action potential triggers a postsynaptic action potential. The approximately 100 billion (10¹¹) neurons in the human brain receive synaptic inputs from many neurons, forming about 10¹⁵ synapses. Synapses cover the surface of dendrites and the cell body (Fig. 11-9A). Some synapses excite the postsynaptic cell by opening ligand-gated cation channels that depolarize the membrane locally. Such small, local changes tend to push the membrane toward threshold for an action potential. Other synapses are inhibitory, hyperpolarizing the postsynaptic membrane locally by opening ligand-gated Cl⁻ channels. These changes are inhibitory, since they drive the membrane potential away from threshold (see Fig. 10-18). From moment to moment, neurons spatially average excitatory and inhibitory stimuli and fire action potentials when the combined effects of these opposing stimuli exceed threshold potential in the proximal part of the axon, called the axon hillock. Both the pattern and frequency of action potentials carry information in the brain.

Transmission at chemical synapses in the CNS depends on cooperation of pumps, carriers, and channels (Fig. 11-9C–D). ATPase pumps maintain concentration gradients of Na⁺, K⁺, and Ca²⁺ across both presynaptic and postsynaptic plasma membranes. Potassium channels establish the resting membrane potential, and voltage-gated K⁺ and Na⁺ channels fire action potentials. Neurotransmitters secreted by the presynaptic cell activate ligand-gated channels that control the postsynaptic membrane potential. Carriers in the presynaptic membrane and adjacent supporting cells terminate transmission by removing neurotransmitter from the synaptic cleft.

Incoming information takes the form of action potentials that arrive at synapses. As in the neuromuscular junction, action potentials open voltage-gated Ca²⁺ channels in the presynaptic membrane. This transient rise in cytoplasmic Ca²⁺ can trigger the fusion of synaptic vesicles with the presynaptic plasma membrane, releasing transmitter, but the probability of successful fusion is lower than at neuromuscular junctions. After vesicle fusion transmitter diffuses rapidly to its receptors in the postsynaptic membrane.

Transmitters activate ligand-gated channels that cause a local, short-lived change in membrane potential, called a postsynaptic potential (PSP). At excitatory synapses, the neurotransmitter glutamate activates receptors (see Fig. 10-11) that open cation channels that depolarize the membrane. However, in contrast to the neuromuscular junction, individual PSPs do not fire action potentials. First, individual PSPs raise the membrane potential only a few millivolts, so they do not bring the postsynaptic membrane to threshold. Second, dendritic and cell body plasma membranes contain few voltage-gated Na⁺ channels. Furthermore, inhibitory synapses on the same cell counteract excitatory synapses by secreting glycine or γ-aminobutyric acid (GABA) to activate Cl⁻ channels that hyperpolarize the membrane, taking it farther from threshold.

Excitatory and inhibitory PSPs spread passively over the postsynaptic membrane and generate an action potential only when their sum at a particular time brings the membrane potential at the axon hillock to threshold (Fig. 11-9A). The axon hillock is located at the base of each axon. This part of the plasma membrane is particularly sensitive to voltage, owing to a high concentration of voltage-gated Na⁺ channels. At threshold, they open, depolarizing the membrane (Fig. 11-6). Delayed-rectifier K⁺ channels then repolarize the membrane in preparation for subsequent action potentials. Each action potential is identical and propagates down the axon.

Because bringing the axon hillock to threshold in the CNS typically requires multiple excitatory postsynaptic potentials, the frequency of the postsynaptic output depends on the intensity of the presynaptic input. In fact, the frequency of postsynaptic action potentials is proportional to the intensity of the presynaptic input above a threshold. This threshold depends on additional voltage-gated K⁺ channels that suppress the firing rate at low levels of stimulation.

Removal of neurotransmitters from the synaptic cleft terminates activation of postsynaptic receptors (Fig. 11-8D). Na⁺K⁺-ATPase pumps provide a Na⁺ gradient to drive symporters that return neurotransmitters to their presynaptic cells. Within the presynaptic cell, a second proton-driven chemiosmotic cycle concentrates transmitter in synaptic vesicles. A V-type, proton-translocating ATPase acidifies the lumen of the synaptic vesicle and establishes the proton electrochemical gradient across the vesicle membrane to drive the antiporter.

Modification of CNS Synapses by Drugs and Disease

The transport systems that retrieve CNS neurotransmitters from the synaptic cleft and repackage them in vesicles determine the duration of synaptic stimulation, so inhibiting these transport processes with drugs prolongs stimulation at particular classes of CNS synapses, with profound effects on brain function and behavior. A plasma membrane dopamine transporter is the main target of cocaine. Cocaine also inhibits transporters for serotonin and norepinephrine. Tricyclic antidepressants inhibit norepinephrine uptake, and other drugs inhibit serotonin uptake. These drugs have dramatic effects on the symptoms of depression as well as a range of milder psychiatric disorders. Millions of people take these drugs, even though the physiological consequences of transporter inhibition are incompletely understood.

Excess stimulation of N-methyl-d-aspartate (NMDA) receptors rapidly kills postsynaptic neurons, most likely owing to the deleterious effects of excess cytoplasmic Ca²⁺. This occurs when glutamate is released from ischemic brain tissue during a stroke caused by compromising the blood supply to a region of the brain. Such damage might also contribute to neuron death in degenerative diseases of the nervous system, such as amyotrophic lateral sclerosis and Alzheimer’s disease.

Acetylcholine secreted by neurons and nicotine from tobacco modulate synaptic transmission in the CNS by activating the same neurotransmitter receptor. In the CNS, acetylcholine acts on presynaptic terminals rather than participating directly in fast synaptic transmission as it does at the neuromuscular junction. Nicotinic acetylcholine receptors in the presynaptic plasma membrane are highly permeable to Ca²⁺, so their stimulation admits Ca²⁺ into the presynaptic terminal. This enhances both the spontaneous release of neurotransmitter and release in response to action potentials. The isoform composition of CNS acetylcholine receptors differs from that of muscles (see Fig. 10-12). Some are homopentamers of α-subunits. Others are heteropentamers of α- and β-subunits. Activation of these ligand-gated channels in different regions of the brain may account for the enhancing effects of nicotine on learning and memory but also for tobacco addiction. Loss of CNS neurons that secrete acetylcholine might contribute to dementia in Alzheimer’s disease.

Modification of CNS Synapses by Use

Memories are thought to be laid down in structural changes that modify the strength or numbers of synapses between neurons in the brain. Particular patterns of stimulation can produce long-term changes that enhance or reduce the efficiency of transmission of various glutamate-mediated synapses (Fig. 11-10). The hippocampus, a region of the vertebrate cerebral cortex that is known to participate in some forms of learning and memory, is favorable for observing a simple form of cellular learning. Intense stimulation of excitatory glutamate synapses (20 pulses over a period of 200 msec) can increase synaptic strength for days or weeks. This is called long-term potentiation (LTP). Conversely, slow, prolonged stimulation of glutamate synapses reduces the response for hours. This is called long-term depression (LTD). The mechanisms of LTP and LTD are under intense investigation, because high-order brain functions, such as learning and memory, depend on changes in the flow of impulses through neural circuits, and the changes in transmission during LTP and LTD occur on an appropriate time scale.

Figure 11-10 mechanism of long-term potentiation of synaptic transmission at excitatory synapses in the hippocampus. A, Prior to long-term potential (LTP), postsynaptic responses to presynaptic action potentials are unreliable and small. B, Some acute responses to vigorous stimulation. C, After induction of LTP, postsynaptic responses are more reliable and larger. AMPA-R and NMDA-R are two classes of glutamate receptors; NO is nitric oxide, a candidate for the retrograde signaling molecule; CAM Kinase II is calcium-calmodulin Kinase II.

Induction of LTP typically involves two types of glutamate receptors (see Fig. 10-11): AMPA receptors and NMDA receptors found in postsynaptic specializations called dendritic spines (Fig. 11-10). AMPA receptors open and close rapidly in response to glutamate. When open, AMPA receptors admit Na⁺ and NMDA receptors admit Ca²⁺ to depolarize the postsynaptic plasma membrane. The slow response of NMDA receptors to glutamate depends on the membrane potential, as partial depolarization is required to displace an extracellular Mg²⁺ ion blocking the channel. This dual dependence on glutamate and membrane potential makes NMDA receptors coincidence detectors, responsive to rapid stimulation or stimulation at nearby excitatory synapses. A synapse with only NMDA receptors is functionally “silent.” Such silent synapses can be aroused when presynaptic release of glutamate is coordinated with sufficient membrane depolarization from neighboring synaptic activation. This results in the insertion of AMPA receptors, waking up the synapse.

In principle, LTP and LTD might alter the efficiency of synaptic transmission by changing glutamate release from the presynaptic cell or responsiveness of the postsynaptic cell to glutamate. In fact, a wide range of experiments suggests that both presynaptic and postsynaptic processes contribute. The best-documented presynaptic change is an increase in the probability that an action potential will stimulate the fusion of a glutamate-containing synaptic vesicle with the plasma membrane. In the resting state, exocytosis of these vesicles is unreliable. LTP increases the probability of exocytosis from less than 0.5 to greater than 0.8. In addition, the postsynaptic side responds more robustly to glutamate for two reasons: a higher concentration of active AMPA receptors with enhanced conductance when open and formation of additional synapses with the stimulating axon.

The mechanisms that bring about these changes are incompletely understood, but the following is well established. LTP depends on stimulation of NMDA receptors and Ca²⁺ entry. Within the postsynaptic dendritic spine, Ca²⁺ activates processes that initiate and maintain LTP. Within seconds, Ca²⁺ binds calmodulin (see Fig. 3-12C) and triggers events that depend on calcium-calmodulin, including activation of protein kinases, such as CAM-kinase II (see Fig. 25-4A). Phosphorylation of AMPA receptors by CAM-kinase II increases their responsiveness to glutamate, perhaps “waking up silent synapses.” AMPA receptors divide their time between the postsynaptic membrane and intracellular recycling endosomes. LTP shifts more AMPA receptors from endosomes to the postsynaptic membrane. No consensus has been reached on the nature of the extracellular messengers that provide feedback to the presynaptic terminal to modify its exocytosis efficiency.

Within minutes, induction of LTP triggers signaling cascades that maintain the increased efficacy, leading to structural changes and increased protein synthesis. These changes may induce dendrites to stabilized existing spines or sprout new filopodia and spines; these are presumed to account for the increased number of synapses observed after an hour or so. Extension of these processes and remodeling of the shape of dendritic spines depend on actin filament assembly (see Fig. 38-8). Growth of axons and formation of new synapses provide mechanisms to generate novel connections in response to use. Over the longer term, the postsynaptic cell initiates gene transcription and protein synthesis, bringing about further changes that stabilize enhanced synaptic transmission.

LTD appears, in many ways, to be the reverse of LTP, with less reliable presynaptic exocytosis and less responsive postsynaptic AMPA receptors. It, too, depends on NMDA receptors, but the biochemical basis for the synaptic changes is even less well understood.

Spontaneous Action Potentials of Pacemaker Cells

Intrinsically excitable pacemaker cells in the sinoatrial node drive rhythmic contractions of the heart (see Fig. 39-19). The membrane potential of these cells drifts spontaneously toward threshold, setting off action potentials about once each second (Fig. 11-11). Cardiac action potentials spread via gap junctions (see Fig. 31-6) from cell to cell throughout the heart, activating contraction in a reproducible pattern.

Figure 11-11 mechanism of spontaneous cardiac pacemaker action potentials. Sequential activation and inactivation of seven different plasma membrane channels account for the time course of action potentials and cytoplasmic Ca²⁺ transients in pacemaker cells of the sinoatrial node. A, Time courses of the fluctuations in membrane potential (orange) and cytoplasmic Ca²⁺ concentration (blue). Colored boxes indicate when the various channels enumerated in part B are open. Kir3.1 and Kir6.2 are inactive under these conditions. B, Channels contributing to pacemaker activity.

Spontaneous action potentials of cardiac pacemaker cells are more complicated than those of nerves. Seven different plasma membrane channels determine their frequency:

Acting together, these channels produce a spontaneous cycle of pacemaker action potentials. At the threshold potential (about -40 mV), voltage-gated Na⁺ channels open synchronously and rapidly depolarize the membrane. As they inactivate, L-type Ca²⁺ channels open, prolonging the depolarization and admitting Ca²⁺; this, in turn, triggers contraction by releasing more Ca²⁺ from internal stores (see Fig. 39-15B). As these Ca²⁺ channels slowly inactivate, delayed-rectifier K⁺ channels open and drive the membrane potential toward E_K, the K⁺ equilibrium potential. As the membrane potential reaches a minimum, delayed-rectifier K⁺ channels inactivate, but the two Kir channels open. In the absence of other channel activity, the membrane potential would remain near E_K, but the nonselective Na⁺/K⁺ channels open and the membrane slowly depolarizes, drifting toward threshold. T-type Ca²⁺ channels contribute to the slow, spontaneous depolarization. At threshold, the cycle repeats.

Mutations in six different human ion channel genes are linked to disorders of cardiac muscle electrophysiology. These inherited diseases are called long-QT syndrome because the interval between the initial de-polarization of the muscle cells and their relaxation is prolonged. This change predisposes the person to abnormal cardiac rhythms that might be fatal.

Regulation of Heart Rate by G Proteins and Phosphorylation

Regulation of pacemaker cells by neurotransmitters secreted by autonomic nerves is an example of the widespread regulation of channels by G proteins and phosphorylation (Fig. 11-12). Neurotransmitters from the two parts of the autonomic nervous system have opposite effects on the frequency of cardiac contraction. The resting rate reflects a compromise in the competition between these two inputs. Acetylcholine from parasympathetic nerves slows the heartbeat, whereas norepinephrine from sympathetic nerves speeds the rate and increases the strength of contraction. These neurotransmitters modify their target channels indirectly by activating two different seven-helix receptors and their associated trimeric G-proteins (see Fig. 25-9).

Norepinephrine increases the heart rate by modulating L-type Ca²⁺ channels. Norepinephrine that binds to plasma membrane β-adrenergic receptors activates trimeric G proteins, which stimulate adenylyl cyclase, the enzyme that makes cyclic adenosine monophosphate (cAMP) (see Fig. 26-2). This second messenger stimulates cyclic AMP–dependent protein kinase (see Fig. 25-3) to phosphorylate cytoplasmic residues of L-type, voltage-gated Ca²⁺ channels in the plasma membrane. Phosphorylated Ca²⁺ channels are more likely to open in response to membrane depolarization than are unphosphorylated channels. Phosphorylation increases the rate at which the membrane potential drifts toward threshold. This increases the frequency of action potentials of pacemaker cells and the heart rate. In a parallel pathway, cAMP stimulates HCN cation channels, which push the membrane potential toward threshold and increase the heart rate.

Acetylcholine released by parasympathetic nerves activates Kir3.1 inward-rectifier K⁺ channels that slow the heartbeat. Acetylcholine binds to different seven-helix receptors, called muscarinic acetylcholine receptors (because they bind muscarine), to distinguish them from the nicotinic acetylcholine receptors. Acetylcholine binding to muscarinic receptors activates a trimeric G protein, different from that activated by norepinephrine. Two G-protein subunits, which make up the Gbg complex, dissociate from the Ga_i subunit and activate Kir3.1/3.4 channels. When open, these channels reduce the rate at which the membrane potential drifts toward threshold. In addition the Ga_i subunit inhibits cyclic AMP production and reduces Ca²⁺ channel phosphorylation. This decreases the probability that Ca²⁺ channels are open, contributing to a lowering of the heart rate. If the energy supply of the heart is compromised, ATP levels fall. This activates Kir6.2 channels, which reduce the rate of spontaneous depolarization and the heart rate until ATP levels are restored.

Regulation of Cardiac Contractility

A set of channels similar to those in the sinoatrial node generate action potentials in cardiac muscle cells and stimulate contraction. Cardiac muscle cells can generate spontaneous action potentials, but they have fewer T-type Ca²⁺ channels and more Kir K⁺ channels, so the rate of spontaneous action potentials is lower than that of pacemaker cells. Except in disease, pacemaker cells drive action potentials throughout the rest of the heart. Sympathetic nervous stimulation, acting through cyclic AMP–dependent protein kinase, strengthens cardiac contraction. Phosphorylated L-type Ca²⁺ channels admit more Ca²⁺ to activate the contractile machinery more fully. The same kinase activates delayed-rectifier K⁺ channels, which prevent activated Ca²⁺ channels from prolonging the action potential. This allows heart muscle cells to keep up with stimuli generated at a higher rate from pacemaker cells. Cyclic AMP–dependent protein kinase also enhances contractility by phosphorylating proteins of the contractile apparatus and smooth endoplasmic reticulum (see Chapter 39).

Therapeutic Effect of Digitalis in Congestive Heart Failure

In congestive heart failure, cardiac contraction fails to produce enough force to maintain adequate circulation of blood. Cardiac glycosides, such as digitalis (from the foxglove plant), ameliorate this common human condition by inhibiting an isoform of Na⁺K⁺ATPase in the plasma membrane of heart cells (Fig. 11-13). Plasma membrane L-type Ca²⁺ channels and endoplasmic reticulum Ca²⁺ release channels activate contraction by transiently increasing cytoplasmic Ca²⁺. Calcium ATPase pumps (see Fig. 8-8) in smooth endoplasmic reticulum clear most of this Ca²⁺ from cytoplasm, but plasma membrane Na⁺/Ca²⁺ antiporters, driven by the plasma membrane Na⁺ gradient, contribute as well.

Figure 11-13 The chemiosmotic cycle that helps to clear Ca²⁺ from the cytoplasm of cardiac muscle cells. Na⁺K⁺-ATPase pumps create a Na⁺ gradient (purple triangle) to drive the Na⁺/Ca²⁺ antiporter to transport Ca²⁺ up its concentration gradient (blue triangle) out of the cell. Cardiac glycosides inhibit the cardiac isoform of Na⁺K⁺-ATPase, raising the concentration of cytoplasmic Ca²⁺ and strengthening cardiac contraction. K⁺ channels (left) allow K⁺ to circulate.

Digitalis and related compounds, such as ouabain, strengthen cardiac contraction indirectly by inhibiting an isoform of Na⁺K⁺-ATPase found in the cardiac plasma membrane. Reduced sodium pump activity lowers the electrochemical gradient of Na⁺ across the plasma membrane and provides less driving force for Na⁺ to exchange for Ca²⁺. The result is a slightly higher steady-state concentration of Ca²⁺ in cytoplasm, which strengthens contraction. A drug that targets Ca²⁺ export directly would be desirable, but thus far, researchers have failed to find a satisfactory inhibitor.

The success of cardiac glycosides depends on their selectivity for the a₂ isoform of the α-subunit of the Na⁺K⁺-ATPase, which is expressed in heart, skeletal muscle, and fat. This a₂ isoform is much more sensitive to cardiac glycosides than is the a₁ isoform. Although the a₂ isoform represents a minority of cardiac sodium pumps, its inhibition lowers the Na⁺ gradient across the heart cell plasma membrane without deleterious effects on other tissues. One reason might be that the Na⁺/Ca²⁺ symporter that works with the sodium pump in the heart is not expressed in skeletal muscle or fat.