Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

the cytoplasmic side of the membrane (Fig. 20-30b) so that

the structure represents the E-1 state of the protein. The

lactose analog is bound in the cavity at a position that is

approximately equidistant from both sides of the mem-

brane, consistent with the model that the lactose binding

site is alternately accessible from each side of the mem-

brane (e.g., Fig. 20-10).Arg, His, and Glu residues that mu-

tational studies have implicated in proton translocation

are located in the vicinity of the lactose binding site.

C. ATP–ADP Translocator

ATP generated in the mitochondrial matrix (its inner com-

partment; Section 1-2A) through oxidative phosphoryla-

tion (Section 22-3C) is largely utilized in the cytosol to

drive such endergonic processes as biosynthesis, active

transport, and muscle contraction.The inner mitochondrial

membrane contains a transmembrane protein that exports

ATP out of the matrix and then imports ADP produced

in the cytosol by ATP hydrolysis. This antiport, the

ATP–ADP translocator (also called the ADP/ATP car-

rier), is electrogenic since it exchanges ADP

3⫺

for ATP

4⫺

.It

is the most abundant member of the mitochondrial carrier

family of proteins that transport a variety of metabolites

through the inner mitochondrial membrane; it constitutes

⬃10% of mitochondrial membrane proteins.

Several natural products inhibit the ATP–ADP translo-

cator. Atractyloside (a poison produced by the Mediter-

ranean thistle Atractylis gummifera that was known to the

ancient Egyptians) and its derivative carboxyatractyloside

(CATR) inhibit the process only from the external surface

of the inner mitochondrial membrane; bongkrekic acid (a

product of the bacterium Pseudomonas cocovenenans)

exerts its effects only on the internal surface.

COOH

OCH

HC C

CCOOH

HC C

COOH

HOOC

–

R = H Atractyloside

R = COOH Carboxyatractyloside

Bongkrekic acid

These differentially acting inhibitors have been valuable

tools in the isolation of the ATP–ADP translocator and in

the elucidation of its mechanism of action.For example,the

translocator has been purified by affinity chromatography

(Section 6-3C) using atractyloside derivatives as affinity

ligands.Atractyloside binding is also a convenient means of

identifying the translocator.

The ATP–ADP translocator, a dimer of identical ⬃300-

residue subunits, has characteristics similar to those of

other transport proteins. It has one binding site for which

ADP and ATP compete. It has two major conformations,

one with its ATP–ADP binding site facing the matrix, and

the other with this site facing outward. In its X-ray struc-

ture in complex with CATR (Fig. 20-31), determined by

Eva Pebay-Peyroula, each subunit’s six transmembrane

helices surround a deep cone-shaped cavity open to the

matrix that is occupied by the CATR.The translocator is an

antiport because it must bind ligand to change from one

conformational state to the other at a physiologically rea-

sonable rate.

The ATP–ADP translocator is not itself an active trans-

port system. However, its electrogenic export of one nega-

tive charge per transport cycle in the direction of ATP

export–ADP import is driven by the membrane potential

difference, ⌬⌿, across the inner mitochondrial membrane

(positive outside).This results in the formation of gradients

in ATP and ADP across the membrane.

5 NEUROTRANSMISSION

In higher animals, the most rapid and complex intercellular

communications are mediated by nerve impulses. A neu-

ron (nerve cell; e.g., Fig. 1-10d) electrically transmits such a

signal along its highly extended length (its axon, which is

commonly over 1 m in larger animals) as a traveling wave

of ionic currents. Signal transmission between neurons, as

well as between neurons and muscles or glands, is usually

chemically mediated by neurotransmitters. In this section

we discuss both the electrical and chemical aspects of nerve

impulse transmission.

A. Voltage-Gated Ion Channels

Ion gradients across cell membranes, as we have seen, are

generated by specific energy-driven pumps (Section 20-3).

These ion gradients are, in turn, discharged through ion

channels such as K

⫹

channels (Section 20-2F). However,

the pumps cannot keep up with the massive fluxes of ions

passing through the channels. Hence ion channels are nor-

mally shut and only open transiently to perform some spe-

cific task for the cell. The opening and closing of ion chan-

nels, a process known as gating, occurs in response to a

variety of stimuli:

1. Ligand-gated channels open in response to extracellu-

lar stimuli.For example,the KcsA K

⫹

channel (Section 20-2F)

opens when the extracellular pH is less than ⬃4 (the X-ray

structure in Fig. 20-12 is that of its closed conformation),

Section 20-5. Neurotransmission 771

JWCL281_c20_744-788.qxd 6/4/10 12:18 PM Page 771

whereas the ligand-gated ion channels in nerve cells open

on extracellularly binding specific neurotransmitters (Sec-

tion 20-5C).

2. Signal-gated channels open on intracellularly bind-

ing a second messenger such as Ca

2⫹

ion or the G

␤␥

subunit

of a heterotrimeric G protein (Section 19-2C).

3. Mechanosensitive channels open under the influence

of stretch, pressure, or displacement. In bacteria, these ap-

pear to function as safety valves to relieve high internal os-

motic pressures that would otherwise burst the cell wall,

whereas in animals they are important in such sensory

functions as touch and hearing.

4. Voltage-gated channels open in response to a change

in membrane potential. Multicellular organisms contain

numerous varieties of voltage-gated channels. For exam-

ple, nerve impulses arise from the sequential opening of

voltage-gated channels along the length of a single nerve

cell (Section 20-5Ba).

a. Voltage Gating in Kv Channels Is Triggered by the

Motion of a Positively Charged Protein Helix

All voltage-gated K

⫹

channels are transmembrane ho-

motetramers, each subunit of which contains an ⬃220-

residue N-terminal cytosolic segment, an ⬃250-residue

transmembrane segment consisting of six helices, S1 to

S6, and an ⬃150-residue C-terminal cytosolic segment

(Fig.20-32).S5 and S6 are homologous to the KcsA channel’s

outer and inner helices (Fig. 20-12a), with their intervening

so-called P loop containing the K

⫹

channel’s TVGYG signa-

ture sequence.

Voltage-gated Na

⫹

channels and Ca

2⫹

channels are

⬃2000-residue monomers that consist of four consecutive

domains, each of which is homologous to the K

⫹

channel

transmembrane domain, separated by often large cytosolic

loops. These domains presumably assume a pseudo-

tetrameric arrangement about a central pore resembling that

of the subunits in voltage-gated K

⫹

channels. This struc-

tural homology suggests that voltage-gated ion channels

772 Chapter 20. Transport Through Membranes

Figure 20-31 X-ray structure of bovine heart ATP–ADP

translocator in complex with carboxyatractyloside. (a) A subunit

of the homodimeric protein, viewed from the plane of the inner

membrane with the intermembrane space above, is drawn in

ribbon form colored in rainbow order from its N-terminus (blue)

to its C-terminus (red) and embedded in its semitransparent

surface diagram.The carboxyatractyloside is drawn in space-

filling form with C cyan, O red, and S yellow. Three cardiolipin

(a)

(

)

molecules that cocrystallized with the protein are drawn in stick

form with C green, O red, and P orange. (b) The molecular

surface as viewed from the intermembrane space and colored

according to its surface charge with blue positive, white neutral,

and red negative. Note the deep positively charged cavity in

which the anionic ATP binds. [Based on an X-ray structure by

Eva Pebay-Peyroula, Université Joseph Fourier, Grenoble,

France. PDBid 2C3E.]

JWCL281_c20_744-788.qxd 7/1/10 7:19 AM Page 772

share a common architecture in which differences in ion se-

lectivity arise from precise stereochemical variations

within the central pore. However, outside of their con-

served transmembrane core, voltage-gated ion channels

with different ion selectivities are highly divergent. For ex-

ample, voltage-gated K

⫹

channels, which are known as Kv

channels, have a conserved ⬃100-residue domain, the so-

called T1 domain (Fig. 20-32), that precedes the transmem-

brane domain and that is absent in other types of voltage-

gated ion channels. The T1 domain confers specificity in

subunit oligomerization: It prevents Kv subunits of differ-

ent subfamilies from coassembling in the same tetramer.

What is the nature of the gating machinery in voltage-

gated ion channels? The X-ray structure of the rat brain Kv

channel named Kv1.2 in its open form, determined by

MacKinnon (Fig.20-33), reveals that its helices S1 through S4

form separate paddlelike domains that extend radially out

from the channel-forming helices, S5 and S6. The 19-residue

S4 helices, which each contain five positively charged

residues (four Arg and one Lys) spaced every ⬃3 residues,

act as voltage sensors.This was shown by covalently linking

a dye whose fluorescence spectrum varies with the polarity

of the environment to any of several residues in S4. Fluo-

rescence measurements on each of these labeled ion chan-

nels revealed that when the membrane potential decreases

(cytosol becomes more negative, which causes the channel

to close), a stretch of at least 7 residues at the N-terminal

end of S4 moves from a position near the extracellular en-

vironment to the center of the membrane. This suggests

that the five positively charged residues on the S4 helix are

drawn toward the cytosol (down in Fig. 20-34a,b) by ⬃15 Å,

pushing down on the S4–S5 linker helix, which in turn

pushes down on the S6 helix, thus collapsing together the

ends of the S6 helices so as to close the cytosolic entrance

to the K

⫹

channel (Fig. 20-34c,d).

b. The Kv Channel Has Two Gates

Electrophysiological measurements indicate that, a few

milliseconds after opening, the Kv channel spontaneously

(without a further change in membrane potential) closes, a

process termed inactivation, and does not reopen until af-

ter the membrane has repolarized (regained its resting

membrane potential). Evidently, the Kv channel contains

two voltage-sensitive gates, one to open the channel on an in-

crease in membrane potential and one to close it a short time

Section 20-5. Neurotransmission 773

Figure 20-32 Secondary structure and membrane orientation

of voltage-gated K

⫹

channels.

Figure 20-33 X-ray structure of the Kv1.2 voltage-gated K

⫹

channel. (a) View along the tetrameric protein’s 4-fold axis from

the extracellular side of the membrane in which it is embedded.

Each of its four identical subunits is colored differently, and

the T1 domain has been omitted for clarity.The fragmented

appearance of the polypeptide chains is due to the high

mobilities of the missing segments.The S5 and S6 helices with

their intervening P loops form the pore for K

⫹

ions (delineated

by green spheres). Helices S1 to S4 form a separate intramem-

brane voltage-sensing domain that associates with the S5 and S6

Inside

Outside

Inactivation ball

S1 S4 S5S3S2

(a)

(b)

helices of the clockwise adjacent subunit. (b) View perpendicular

to that in Part a with the extracellular side of the membrane

above. The pore and voltage sensing domains span a distance of

⬃30 Å, the thickness of the membrane’s hydrophobic core. The

T1 domain, which occupies the cytoplasm, forms the vestibule of

the transmembrane K

⫹

channel.The four large openings

between the T1 domain and the K

⫹

channel are the portals

through which K

⫹

ions enter the K

⫹

channel. [Based on an X-ray

structure by Roderick MacKinnon, Rockefeller University.

PDBid 2A79.]

JWCL281_c20_744-788.qxd 7/1/10 12:22 PM Page 773

later, a phenomenon that has important consequences for

the transmission of nerve impulses (Section 20-5B). In fact,

minute “gating currents” arising from the movements of

these positively charged gates in opening and closing can

be detected (electrical current is the movement of charge)

if the much larger currents of K

⫹

ions through the mem-

brane are first blocked by plugging the Kv channel from its

cytosolic side by high concentrations of Cs

⫹

or tetraethyl-

ammonium ions (which are too large to pass through the

⫹

pore but apparently become stuck within it).

c. Inactivation Occurs through the Insertion of the Kv

Channel’s N-Terminal Peptide into Its Central Pore

The inactivation of the Kv channel is abolished by pro-

teolytically excising its N-terminal ⬃20-residue segment

(its inactivation peptide), which NMR studies indicate

forms a ball-like structure. However, when chemically syn-

thesized inactivation peptide is injected into the cytosol,

the truncated Kv channel inactivates at a rate proportional

to the concentration of inactivation peptide. This suggests

that inactivation occurs when the inactivation ball swings

around at the end of the ⬃65-residue peptide linking it to

the T1 domain so as bind to the open K

⫹

pore in a way that

physically blocks the passage of K

⫹

ions—the so-called

ball-and-chain mechanism. Indeed, the time the Kv chan-

nel stays open varies with the length of this “chain,” as mu-

tationally adjusted.

Where is the inactivation peptide’s binding site? Muta-

tional analysis by MacKinnon revealed that the hydropho-

bic residues lining the Kv channel’s internal pore and cen-

tral cavity form the receptor site for the inactivation

peptide. Since the ⬃6-Å-diameter cytosolic entrance to the

internal pore is too narrow to admit the ball, the ball pep-

tide must unfold in order to enter the internal pore. The

first 10 residues of ball peptides are predominantly hy-

drophobic, whereas the succeeding 10 residues are largely

hydrophilic and contain several basic residues. It therefore

appears that inactivation occurs through the binding of the

N-terminus of the fully extended inactivation peptide in-

side the internal pore via hydrophobic interactions, an as-

sociation that is augmented by the binding of the basic

residues in the ball peptide’s C-terminal segment to the

acidic residues lining the entrance of the internal pore.

Thus, the inactivation peptide acts more like a snake than a

ball and chain.

How does the inactivation peptide gain access to the inter-

nal pore, which appears to be covered by the T1 tetramer?

The passage through the center of the T1 tetramer, as seen

in its X-ray structure (Fig. 20-33b), is too narrow to permit

the passage of the inactivation peptide. Moreover, Christo-

pher Miller eliminated the possibility that the individual T1

domains in a tetramer can ever separate far enough to ad-

mit the inactivation peptide by showing that cross-linking

adjacent T1 domains by genetically engineered disulfide

bonds (whose positions were selected by referring to the

T1 X-ray structure) does not significantly affect the Kv

channel’s gating properties. This strongly suggests that the

inactivation peptide gains access to the bottom of the

transmembrane pore through lateral windows between the

transmembrane and T1 domains, whose sides are formed

by the ⬃35-residue peptide segment linking these domains.

Presumably, K

⫹

ions pass through these same windows

when the Kv channel is open.

A Kv channel engineered so that only one subunit has

an inactivation peptide still inactivates but at one-fourth

the rate of normal Kv channels. Apparently, any of the nor-

mal Kv channel’s four inactivation peptides can block the

774 Chapter 20. Transport Through Membranes

(a)

(c)

(d)

(b)

Figure 20-34 Proposed mechanism of voltage gating in the

Kv1.2 voltage-gated K

⫹

channel. (a) The voltage-sensing domain

in the X-ray structure of Kv1.2 in its open conformation as

viewed from the pore parallel to the plane of the membrane with

the extracellular solution above. The polypeptide chain is drawn

in ribbon form with helices S1 through S4 white and the S4–S5

linker helix orange. Side chains are shown in ball-and-stick form

with those of the Arg and Lys residues of S4, the so-called gating

charges, cyan, those of Asp and Glu in the so-called external and

internal clusters red, and that of Phe 233, which separates the

external and internal clusters, green. Note that the gating charges

extend upward into the extracellular solution and associate with

the external cluster. (b) The voltage-sensing domain in the

hypothetical closed conformation displayed as in Part a. S4,

together with S3, has moved downward relative to S1 and S2

toward the now more negatively charged cytosol, thereby

pivoting the S4-S5 helix in the direction of the white arrow in

Part a. Here the gating charges extend downward toward the

cytosol and associate with the internal cluster. (c) The S4–S6

linkers and the K

⫹

pore assembly, which includes helices S5 and

S6, in the X-ray structure of Kv1.2 as viewed in the opposite

direction from Part a.The S6 helices are blue, the S4–S5 linkers,

which contact S6, are orange, and the remainder of the pore

assembly is green. (d) Hypothetical model of the S4–S5 linker in

the closed conformation, displayed as in Part c, with the structure

of the K

⫹

pore assembly based on the X-ray structure of the

KscA channel (Fig. 20-12a). [Courtesy of Roderick MacKinnon,

Rockefeller University. PDBid 2R9R.]

JWCL281_c20_744-788.qxd 7/1/10 7:58 AM Page 774

channel and it is simply a matter of chance as to which one

does so. In contrast, voltage-gated Na

⫹

channels have only

a single inactivation peptide, which is located on the seg-

ment linking the Na

⫹

channel’s third and fourth homolo-

gous transmembrane domains. Consequently, a genetically

engineered cut of the peptide chain in this region abolishes

⫹

channel inactivation.

B. Action Potentials

Neurons, like other cells, generate ionic gradients across

their plasma membranes through the actions of the corre-

sponding ion-specific pumps. In particular, an (Na

⫹

–K

⫹

)–

ATPase (Section 20-3A) pumps K

⫹

into and Na

⫹

out of the

neuron to yield intracellular and extracellular concentra-

tions of these ions similar to those listed in Table 20-3. The

consequent membrane potential, ⌬⌿, across a cell mem-

brane is described by the Goldman equation, an extension

of Eq. [20.3] that explicitly takes into account the various

ions’ different membrane permeabilities:

[20.8]

Here, C and A represent cations and anions, respectively,

and, for the sake of simplicity, we have made the physiolog-

ically reasonable assumption that only monovalent ions

have significant concentrations.The quantities P

and P

, the

respective permeability coefficients for the various cations

and anions, are indicative of how readily the corresponding

ions traverse the membrane (each is equal to the correspon-

ding ion’s diffusion coefficient through the membrane

divided by the membrane’s thickness; Section 20-2A). Note

that Eq. [20.8] reduces to Eq. [20.3] if the permeability coef-

ficients of all mobile ions are assumed to be equal.

Applying Eq. [20.8] to the data in Table 20-3 and assum-

ing a temperature of 25°C yields ⌬⌿ ⫽ ⫺83 mV (negative

inside), which is in good agreement with experimentally

measured membrane potentials for mammalian cells. This

value is somewhat greater than the K

⫹

equilibrium potential,

the value of ⌬⌿ ⫽ –91 mV obtained assuming the membrane

is permeable to only K

⫹

ions .The mem-

brane potential is generated by a surprisingly small imbal-

ance in the ionic distribution across the membrane: Only

⬃1 ion pair per million is separated by the membrane with

⫹

⫽ P

⫺

⫽ 0)

¢⌿⫽

[C(out)] ⫹

[A(in)]

[C(in)] ⫹

[A(out)]

the anion going to the cytosolic side and the cation going to

the external side.The resulting electric field is, nevertheless,

enormous by macroscopic standards: Assuming a typical

membrane thickness of 50 Å, it is nearly 170,000 V ⴢ cm

⫺1

a. Nerve Impulses Are Propagated by

Action Potentials

A nerve impulse consists of a wave of transient mem-

brane depolarization known as an action potential that

passes along a nerve cell. A microelectrode implanted in an

axon will record that during the first ⬃0.5 ms of an action

potential, ⌬⌿ increases from its resting potential of

around –60 mV to about ⬃30 mV (Fig. 20-35a).This depo-

larization is followed by a nearly as rapid repolarization

Section 20-5. Neurotransmission 775

Table 20-3 Ionic Concentrations and Membrane

Permeability Coefficients in Mammals

Cell Blood Permeability

Ion (mM)(mM) Coefficient (cm ⴢ s

⫺1

)

⫹

139 4 5 ⫻ 10

⫺7

⫹

12 145 5 ⫻ 10

⫺9

⫺

4 116 1 ⫻ 10

⫺8

⫺a

138 9 0

⫺

represents macromolecules that are negatively charged under

physiological conditions.

Source: Darnell, J., Lodish, H., and Baltimore, D., Molecular Cell Biology,

pp. 618 and 725, Scientific American Books (1986).

Figure 20-35 Time course of an action potential. (a) The axon

membrane undergoes rapid depolarization, followed by a nearly

as rapid hyperpolarization and then a slow recovery to its resting

potential. (b) The depolarization is caused by a transient increase

in Na

⫹

permeability (conductance), whereas the hyperpolariza-

tion results from a more prolonged increase in K

⫹

permeability

that begins a fraction of a millisecond later.The unit of conduc-

tance, 1 mho ⫽ 1 ohm

⫺1

. [After Hodgkin, A.L. and Huxley, A.F.,

J. Physiol. 117, 530 (1952).]

equilibrium potential

Action potential

Resting potential

equilibrium potential

Hyperpolarization

Depolarization

+60

+40

+20

–20

–40

–60

–80

Membrane potential (mV)

0 123 4

Time (ms)

0 123 4

Time (ms)

Ionic permeabilities (mmho • cm

–2

)

permeability

(a)

(b)

JWCL281_c20_744-788.qxd 6/4/10 1:39 PM Page 775

past the resting potential to the K

⫹

equilibrium potential

(hyperpolarization) and then a slower recovery to the

resting potential. What is the origin of this complicated

electrical behavior? In 1952, Alan Hodgkin and Andrew

Huxley demonstrated that the action potential results

from a transient increase in the membrane’s permeability

to followed, within a fraction of a millisec-

ond, by a transient increase in its permeability to K

⫹

(

Fig. 20-35b).

The ion-specific permeability changes that characterize

an action potential result from the presence of Na

⫹

- and K

⫹

specific voltage-gated channels. As a nerve impulse reaches

a given patch of nerve cell membrane, the increased mem-

brane potential induces the transient opening of the Na

⫹

channels, so that Na

⫹

ions diffuse into the nerve cell at the

rate of ⬃6000 ions ⴢ ms

⫺1

per channel.This increase in

causes to increase (Eq. [20.3]), which, in turn, induces

more Na

⫹

channels to open, etc., leading to an explosive

entry of Na

⫹

into the cell.Yet, before this process can equil-

ibrate at its Na

⫹

equilibrium potential of around ⬃60 mV,

the K

⫹

channels open ( increases) while the Na

⫹

chan-

nels close (inactivate; returns to its resting value).

therefore reverses sign and overshoots its resting potential

to approach its K

⫹

equilibrium value. Eventually the K

⫹

channels also inactivate and the membrane patch regains

its resting potential. The Na

⫹

channels, which remain open

only 0.5 to 1.0 ms, do not reopen until the membrane has

returned to its resting state, thereby limiting the axon’s

firing rate.

¢⌿P

⫹

¢⌿

⫹

;

⫹

)

An action potential is triggered by an ⬃20-mV rise in

⌬⌿ to about ⫺40 mV. Action potentials therefore propagate

along an axon because the initially rising value of ⌬⌿ in a

given patch of axonal membrane triggers the action poten-

tial in an adjacent membrane patch that does so in an adja-

cent membrane patch, etc. (Fig. 20-36). The nerve impulse

is thereby continuously amplified so that its signal ampli-

tude remains constant along the length of the axon (in

contrast, an electrical impulse traveling down a wire dissi-

pates as a consequence of resistive and capacitive ef-

fects). Note, however, that since the relative ion imbal-

ance responsible for the resting membrane potential is

small, only a tiny fraction of a nerve cell’s Na

⫹

–K

⫹

gradi-

ent is discharged by a single nerve impulse (only one K

⫹

ion per 3000–300,000 in the cytosol is exchanged for ex-

tracellular Na

⫹

, as indicated by measurements with ra-

dioactive Na

⫹

). An axon can therefore transmit a nerve

impulse every few milliseconds without letup. This capac-

ity to fire rapidly is an essential feature of neuronal com-

munications: Since nerve impulses all have the same ampli-

tude, the magnitude of a stimulus is conveyed by the rate at

which a nerve fires.

b. The Voltage-Gated Na

⫹

Channel Is the

Target of Numerous Neurotoxins

Neurotoxins have proved to be invaluable tools for dis-

secting the various mechanistic aspects of neurotransmis-

sion. Many neurotoxins, as we shall see, interfere with the

action of neuronal voltage-gated Na

⫹

channels but,curiously,

776 Chapter 20. Transport Through Membranes

Figure 20-36 Action potential propagation along an axon.

Membrane depolarization at the leading edge of an action

potential triggers an action potential at the immediately

downstream portion of the axon membrane by inducing the

opening of its voltage-gated Na

⫹

channels.As the depolarization

wave moves farther downstream, the Na

⫹

channels close and the

⫹

channels open to hyperpolarize the membrane. After a brief

refractory period, during which the K

⫹

channels close and the

hyperpolarized membrane recovers its resting potential, a second

0 5 10 15 20 25 30

Membrane potential (mV)

+40

–40

–80

+++++++ +++

Minimum

refractory period

10 ms

+++++++++++ ++++++–– ––

+++++++ ++++++++++++++ ++++++–– ––

++––––––– –––––––––––––– ––––––++

impulse can follow. The indicated impulse propagation speed is

that measured in the giant axon of the squid, which, because of

its extraordinary width (⬃1 mm), is a favorite experimental

subject of neurophysiologists. Note that the action potential in

this figure appears backward from that in Fig. 20-35 because this

figure shows the distribution of the membrane potential along an

axon at an instant in time, whereas Fig. 20-35 shows the

membrane potential’s variation with time at a fixed point on the

axon.

JWCL281_c20_744-788.qxd 3/17/10 1:49 PM Page 776

few are known that affect voltage-gated K

⫹

channels.

Tetrodotoxin,

a paralytic poison of enormous potency, which occurs

mainly in the skin, ovaries, liver,and intestines of the puffer

fish (known as fugu in Japan, where it is a delicacy that may

be prepared only by chefs certified for their knowledge of

puffer fish anatomy), acts by specifically blocking the Na

⫹

channel.The Na

⫹

channel is similarly blocked by saxitoxin,

a product of marine dinoflagellates (a type of plankton

known as the “red tide”) that is concentrated by filter-feeding

shellfish to such an extent that a small mussel can contain

sufficient saxitoxin to kill 50 people. Both of these neuro-

toxins have a cationic guanidino group, and both are effec-

tive only when applied to the external surface of a neuron

(their injection into the cytosol elicits no response). It is

therefore thought that these toxins specifically interact

with an anionic carboxylate group located at the mouth of

the Na

⫹

channel on its extracellular side.

Batrachotoxin,

Saxitoxin

–

Tetrodotoxin

arrow-poison frog, Phyllobates aurotaenia, is the most po-

tent known venom (2 ␮g ⴢ kg

⫺1

body weight is 50%

lethal in mice). This substance also specifically binds to

the voltage-gated Na

⫹

channel but, in contrast to the ac-

tions of tetrodotoxin and saxitoxin, renders the axonal

membrane highly permeable to Na

⫹

. Indeed, batra-

chotoxin-induced axonal depolarization is reversed by

tetrodotoxin.The observation that the repeated electrical

stimulation of a neuron enhances the action of batra-

chotoxin indicates that this toxin binds to the Na

⫹

chan-

nel in its open state.

Venoms from American scorpions contain families of

60- to 70-residue protein neurotoxins that also act to depo-

larize neurons by binding to their Na

⫹

channels (the differ-

ent neurotoxins in the same venom appear to be special-

ized for binding to the Na

⫹

channels in the various species

the scorpion is likely to encounter). Scorpion toxins and

tetrodotoxin do not, however, compete with each other for

binding to the Na

⫹

channel and therefore must bind at sep-

arate sites.

c. Nerve Impulse Velocity Is Increased

by Myelination

The axons of the larger vertebrate neurons are sheathed

with myelin, a biological “electrical insulating tape” that is

wrapped about the axon (Fig. 20-37a) so as to electrically

isolate it from the extracellular medium. Impulses in myeli-

nated nerves propagate with velocities of up to 100 m ⴢ s

⫺1

whereas those in unmyelinated nerves are no faster than

10 m ⴢ s

⫺1

(imagine the coordination difficulties that, say, a

giraffe would have if it had to rely on only unmyelinated

nerves).

How does myelination increase the velocity of nerve

impulses? Myelin sheaths are interrupted every millimeter

or so along the axon by narrow unmyelinated gaps known

as nodes of Ranvier (Fig. 20-37b), where the axon contacts

the extracellular medium. Binding studies using radioac-

tive tetrodotoxin indicate that the voltage-gated Na

⫹

channels of unmyelinated axons have rather sparse al-

though uniform distributions in the axonal membrane of

⬃20 channels ⴢ ␮m

⫺2

. In contrast, the Na

⫹

channels of

myelinated axons occur only at the nodes of Ranvier,

where they are concentrated with a density of ⬃10

chan-

nels ⴢ ␮m

⫺2

. The action potential of a myelinated axon

evidently hops between these nodes, a process named

saltatory conduction (Latin: saltare, to jump). Nerve im-

pulse transmission between the nodes must therefore oc-

cur by the passive conduction of an ionic current, a mech-

anism that is inherently much faster than the continuous

propagation of an action potential but that is also dissipa-

tive. The nodes act as amplification stations to maintain

the intensity of the electrical impulse as it travels down the

axon.Without the myelin insulation, the electrical impulse

would become too attenuated through transmembrane

ion leakage and capacitive effects to trigger an action po-

tential at the next node. In fact, multiple sclerosis, an au-

toimmune disease that demyelinates nerve fibers in the

brain and spinal cord, results in serious although rarely

fatal neurological deficiencies.

Section 20-5. Neurotransmission 777

Batrachotoxin

a steroidal alkaloid secreted by the skin of a Colombian

JWCL281_c20_744-788.qxd 7/1/10 7:20 AM Page 777

C. Neurotransmitters and Their Receptors

The junctions at which neurons pass signals to other neu-

rons, muscles, or glands are called synapses. In electrical

synapses, which are specialized for rapid signal transmis-

sion, the cells are separated by a gap, the synaptic cleft, of

only 20 Å,which is spanned by gap junctions (Section 12-3F).

Hence, an action potential arriving at the presynaptic side

of the cleft can sufficiently depolarize the postsynaptic

membrane to trigger its action potential directly. How-

ever, the ⬎200-Å gap of most synapses is too large a dis-

tance for such direct electrical coupling. In these chemical

synapses, the arriving action potential triggers the release

from the presynaptic neuron of a specific substance

known as a neurotransmitter, which diffuses across the

cleft and binds to its corresponding receptors on the post-

synaptic membrane. In excitatory synapses, neurotrans-

mitter binding induces membrane depolarization, thereby

triggering an action potential on the postsynaptic mem-

brane. Conversely, neurotransmitter binding in inhibitory

synapses alters postsynaptic membrane permeability so as

to inhibit an action potential and thus attenuate excitatory

signals. What is the mechanism through which an arriving

action potential stimulates the release of a neurotransmit-

ter, and by what means does its binding to a receptor alter

the postsynaptic membrane’s permeability? To answer

these questions let us consider the workings of cholinergic

synapses; that is, synapses that use acetylcholine (ACh) as

a neurotransmitter:

Two types of cholinergic synapses are known:

1. Those containing nicotinic receptors (receptors that

respond to nicotine).

2. Those containing muscarinic receptors (receptors

that respond to muscarine, an alkaloid produced by the

poisonous mushroom Amanita muscaria).

Nicotine

Acetylcholine (ACh)

Muscarine

N(CH

)

N(CH

)

778 Chapter 20. Transport Through Membranes

Figure 20-37 Myelination. (a) An electron micrograph of

myelinated nerve fibers in cross section.The myelin sheath

surrounding an axon is the plasma membrane of a Schwann cell,

which, as it spirally grows around an axon, extrudes its cytoplasm

from between the layers.The resulting double bilayer, which

makes between 10 and 150 turns about the axon, is a good

electrical insulator because of its particularly high (79%) lipid

content. [Courtesy of Cedric Raine,Albert Einstein College of

Medicine of Yeshiva University.] (b) A schematic diagram of a

myelinated axon in longitudinal section, indicating that in the

nodes of Ranvier (the relatively short gaps between adjacent

myelinating cells), the axonal membrane is in contact with the

extracellular medium.A depolarization generated by an action

potential at one node hops, via ionic conduction, down the

myelinated axon (red arrows), to the neighboring node, where it

induces a new action potential. Nerve impulses in myelinated

axons are therefore transmitted by saltatory conduction.

+++

–––

+++

–––

+++

–––

+++

+++ +++

–––

+++

–––

+++

–––

+++

–––

Myelin sheath

Axon

Node of

Ranvier

(b)

(a)

JWCL281_c20_744-788.qxd 6/4/10 1:23 PM Page 778

In what follows, we shall focus on cholinergic synapses con-

taining nicotinic receptors since this best characterized type

of synapse occurs at all excitatory neuromuscular junctions

in vertebrates and at numerous sites in the nervous system.

a. Electric Organs of Electric Rays Are Rich

Sources of Cholinergic Synapses

The study of synaptic function has been greatly facili-

tated by the discovery that the homogenization of nerve

tissue causes its presynaptic endings to pinch off and reseal

to form synaptosomes (Section 12-4Da). The use of synap-

tosomes, which can be readily isolated by density gradient

ultracentrifugation, has the advantage that they can be ma-

nipulated and analyzed without interference by other neu-

ronal components.

The richest known source of cholinergic synapses is the

electric organs of the freshwater electric eel Electrophorus

electricus and saltwater electric rays of the genus Torpedo.

Electric organs, which these organisms use to stun or kill

their prey, consist of stacks of ⬃5000 thin flat cells called

electroplaques that begin their development as muscle

cells but ultimately lose their contractile apparatus. One

side of an electroplaque is richly innervated and has high

electrical resistance, whereas its opposite side lacks inner-

vation and has low electrical resistance. Both sides main-

tain a resting membrane potential of around ⫺90 mV. On

neuronal stimulation, all the innervated membranes in a

stack of electroplaques simultaneously depolarize to a

membrane potential of around ⬃40 mV, yielding a poten-

tial difference across each cell of 130 mV (Fig. 20-38). Since

the 5000 electroplaques in a stack are “wired” in series like

the batteries in a flashlight, the total potential difference

across the stack is ⬃5000 ⫻ 0.130 V ⫽ 650 V, enough to

kill a human being.

b. Acetylcholine Is Released by the Ca

2ⴙ

-Triggered

Exocytosis of Synaptic Vesicles

ACh is synthesized near the presynaptic end of a neuron

by the transfer of an acetyl group from acetyl-CoA [the

structure of coenzyme A (CoA) is given in Fig. 21-2] to

choline in a reaction catalyzed by choline acetyltransferase.

Much of this ACh is sequestered in ⬃400-Å-diameter

membrane-enveloped synaptic vesicles, which typically

contain ⬃10

ACh molecules each.

The arrival of an action potential at the presynaptic mem-

brane triggers the opening of voltage-gated Ca

2ⴙ

channels,

which transiently raises the local [Ca

2⫹

] from its resting level

of 0.1 ␮M to 10 to 100 ␮M. The resulting influx of extracel-

lular Ca

2⫹

, in turn, stimulates the exocytosis of the synaptic

vesicles in the vicinity of the Ca

2⫹

channel so that they re-

lease their packets of ACh into the synaptic cleft (Fig. 12-73).

The mechanism by which synaptic vesicles fuse with the

presynaptic membrane is discussed in Section 12-4D.

The mechanism through which Ca

2⫹

induces synaptic

vesicle exocytosis is beginning to come into focus. The ma-

jor Ca

2⫹

-sensing protein appears to be synaptotagmin I, a

protein with a single helix passing through the synaptic

vesicle membrane, whose cytosolic domain contains four

2⫹

binding sites.At resting levels of Ca

2⫹

, synaptotagmin

I binds to the Q-SNARE syntaxin (Section 12-4Db) so as

to block its binding to the R-SNARE synaptobrevin and

the Q-SNARE SNAP25,thereby preventing vesicle fusion.

However, on binding Ca

2⫹

, synaptotagmin I releases syn-

taxin, permitting vesicle fusion to commence.

Once triggered, the fusion of the synaptic vesicles with

the presynaptic membrane occurs very rapidly (in ⬍0.3 ms)

Acetyl-CoA Choline

Acetylcholine

C CoA +

HO CH

N(CH

)

N(CH

)

HSH

choline

acetyltransferase

Section 20-5. Neurotransmission 779

Figure 20-38 The simultaneous depolarization (red, right) of

the innervated membranes in a stack of electroplaques “wired”

in series results in a large voltage difference between the two

++++++

+++++++

–––––––

Motor

neuron

Electroplaques

Resting state Depolarized

130 mV

+40 mV

–90 mV

0 mV

–90 mV

+++++++

–––––––

–

+++++++

+++++++ +++++++

–––––––

––––––

++++++

–––––––

+++++++

–––––––

+++++++

–

+++++++

–––––––

+++++++

–––––––

––––––

ends of the stack. This is because the total voltage across the

stack is the sum of the voltages generated by each of its numer-

ous electroplaques.

JWCL281_c20_744-788.qxd 3/17/10 1:49 PM Page 779

because many synaptic vesicles are already docked

with the presynaptic membrane. Each Ca

2⫹

pulse triggers

the exocytosis of ⬃10% of these docked vesicles. How-

ever, they are rapidly replaced because most of the re-

maining synaptic vesicles are held in reserve in a so-called

active zone within 20 nm of the presynaptic membrane.

Vesicles are held in the active zone by a fibrous phospho-

protein named synapsin I that also binds the cytoskeletal

proteins actin and spectrin (Section 12-3Db). Synapsin I

is a substrate for calmodulin-dependent protein kinase

(Section 18-3Ce), so that a rise in [Ca

2⫹

] causes its phos-

phorylation.This apparently releases the synaptic vesicles

from the active zone, thereby permitting them to dock to

the presynaptic membrane in preparation for exocytosis.

The 20-nm distance between the active zone and the

presynaptic membrane may be close enough for Q- and R-

SNAREs to initiate the formation of their coiled coil com-

plex (Fig. 12-74), which may thereby facilitate the docking

process.

The black widow spider takes advantage of this system:

Its highly neurotoxic venom protein, ␣-latrotoxin (130 kD),

causes massive release of ACh at the neuromuscular junc-

tion, in part by forming homotetrameric transmembrane

channels through the presynaptic membrane that act as

2⫹

ionophores. In contrast, botulinus toxin, as we have

seen (Section 12-4Dd), interferes with the exocytosis of

synaptic vesicles by proteolytically cleaving specific

SNARE proteins, thereby preventing ACh release.

Exocytosed synaptic vesicle proteins are rapidly recov-

ered from the presynaptic membrane via endocytosis

mainly in clathrin-coated vesicles (Section 12-4C). How-

ever, once the resulting endocytotic vesicles lose their

clathrin coats, they do not fuse with endosomes, as is usu-

ally the case (Fig. 12-91). Rather, they are immediately re-

filled with ACh by an H

⫹

–ACh antiport, which is driven by

the protons pumped into the vesicle by a V-type ATPase

(Section 20-3), and then translocated to the active zone.

This rapid recycling of the synaptic vesicles (which takes

⬍1 min) permits neurons to fire continuously at a rate of

⬃50 times per second.

c. The Acetylcholine Receptor Is a

Ligand-Gated Cation Channel

After its release at the presynaptic membrane, acetyl-

choline quickly diffuses across the synaptic cleft to the

postsynaptice membrane, where it binds to the 290-kD

acetylcholine receptor (AChR), an ␣

␤␥␦ transmembrane

glycoprotein, whose four different ⬃490-residue subunits

are homologous. The binding of two ACh molecules to the

AChR allosterically induces the opening of a channel

through the AChR that permits Na

⫹

ions to diffuse into the

cell at a peak rate of ⬃30,000 ions per millisecond. The

780 Chapter 20. Transport Through Membranes

Figure 20-39 Electron crystal structure of the nicotinic

acetylcholine receptor from the electric ray Torpedo marmorata.

(a) View from the synaptic cleft. The extracellular domains of

each type of subunit are differently colored and their remaining

portions are gray. The side chains of ␣Trp 149, which mark the

ACh-binding sites, are drawn in stick form (gold). (b) View

(b)

(a)

parallel to the membrane with the synaptic cleft above. Only the

front two subunits are highlighted in color.The horizontal gray

bars delineate the position of the postsynaptic membrane.

[Courtesy of Nigel Unwin, MRC Laboratory of Molecular

Biology, Cambridge, U.K. PDBid 2BG9.]

JWCL281_c20_744-788.qxd 3/17/10 1:49 PM Page 780