Mechanisms of ion channel activation in olfactory receptor cells of the spiny lobster, Panulirus argus


Material Information

Mechanisms of ion channel activation in olfactory receptor cells of the spiny lobster, Panulirus argus
Physical Description:
x, 176 leaves : ill. ; 28 cm.
McClintock, Timothy S., 1959-
Publication Date:


Subjects / Keywords:
Spiny lobsters -- Physiology   ( lcsh )
Ion channels   ( lcsh )
Cell receptors   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1989.
Includes bibliographical references (leaves 164-175).
Statement of Responsibility:
by Timothy S. McClintock.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001479601
notis - AGZ1559
oclc - 20967466
System ID:

Full Text







Copyright 1989


Timothy S. McClintock



First, I thank my wife, Kathy, for her love, support,

encouragement, and understanding which helped to make these last

four years such fun. Without complaint, she made many sacrifices

which made this dissertation possible.

With deepest gratitude, I thank my advisor, Dr. Barry Ache, for

the innumerable ways in which he contributed to my development as a

scientist. He provided me with many wonderful opportunities and

challenged me with responsibility. The best and most enjoyable part

of these last few years has been working with this kind and patient


I thank Dr. Peter Anderson and Dr. Barbara Battelle for

teaching me techniques, loaning equipment, and for innumerable

pieces of advice. I also thank Dr. Bill Carr for much advice, but I

am most grateful to him for infecting me with some of his famous

enthusiasm on a few occasions when I really needed it. I thank Dr.

Michael Greenberg for discussions of our common interest in cellular

mechanisms of action, and for sharing his knowledge of the


I thank the entire faculty and staff of the Whitney Laboratory

for making my stay so pleasant. All the individual laboratories

shared both their equipment and expertise at some point during my

dissertation work, and I am grateful for how much they taught me. I

am especially grateful to Dr. Sam Edwards, Dr. Henry Trapido-

Rosenthal, Dr. Rick Gleeson, Dr. David Price, Dr. John Schell, John

Young, and Ann Wishart in this regard. I thank Shirley Metz for

bailing me out of a sea of red tape on a few occasions. Thanks also

go to Lynn Milstead and Jim Netherton for converting my crude

illustrations into beautiful figures and slides.

I thank the Department of Zoology for helping to support my

training in electrophysiological techniques at The Marine Biological

Laboratory. I thank Dr. David Evans and Dr. Frank Nordlie for their

advice and encouragement. I thank Carol Binello and Grace Kiltie

for the many times they helped me cope with the university's









The Olfactory Organ
Statement of Purpose

of the Spiny Lobster

S 1


Introduction .
Methods .
Preparation of Isolated Olfactory Receptor
Receptor Cell Somata. .
Voltage-clamp Recording .
Solutions. .
Results .
Total Membrane Currents .
Na* Current .
Ca-- Current .
Ca*-activated K" Current. .
Delayed Rectifier K- Current .
Discussion. .


Introduction .
Methods .
Preparation of Receptor Cells.
Patch Clamp Recording.
Solutions. .
Results .

A Voltage-dependent
A Ca*-activated K

Cation Channe

S 14

S 16
S 17
S 18


1 .




A Voltage-dependent K- Channel
A Steady State Cl- Channel
Discussion. .
The Nonselective Cation Channel
K' Channels .
Cl- Channel
Significance for Olfaction


Introduction .
Methods .
Preparation .
Intracellular Recording
Solutions and Chemicals
Results .
Intracellular Recording
Discussion. .




Introduction .
Methods .
Results .
Discussion. .





. 110

. 111









. 176


ADP adenosine 5'-diphosphate

AMP = adenosine 5'-monophosphate

ATP = adenosine 5'-triphosphate

8-bromo-cAMP = 8-bromo-cyclic adenosine 3',5'-monophosphate

8-bromo-cGMP = 8-bromo-cyclic guanosine 3',5'-monophosphate

cAMP = cyclic adenosine 3',5'-monophosphate

cGMP cyclic guanosine 3',5'-monophosphate

CNS central nervous system

EGTA ethylene glycol bis-N,N,N',N'-tetraacetic acid

fc = corner frequency

GDP-3-S guanosine 5'-0-(2-thiodiphosphate)

GTP = guanosine 5'-triphosphate

GTP-T-S guanosine 5'-0-(3-thiotriphosphate)

HA histamine

HEPES N-2-hydroxyethyl piperazine-N'-2 ethane sulfonic acid

I current

IBMX l-isobutyl-3-methylxanthine

Po probability of an ion channel being in the open state

R resistance

TEA tetraethylammonium

TTX tetrodotoxin

V voltage


Abstract of Dissertation Presented to the Graduate
School of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy



Timothy S. McClintock

May, 1989

Chairman: Barry W. Ache
Major Department: Zoology

As in any cell, ion channels underlie the active electrical

properties of the lobster olfactory receptor cell. Studying the

mechanisms of activation of these ion channels is a step toward

understanding how the olfactory receptor cell performs its


Somata isolated from olfactory receptor cells contained three

voltage-dependent currents: a sustained Ca" current, a Ca-'-

activated K* current, and a delayed rectifier K- current. A TTX-

sensitive Na* current occurred predominantly in the axon. Four ion

channels were described: a 35 pS voltage-independent Cl- channel, a

9.7 pS delayed rectifier K- channel, a 215 pS Ca-*-activated K-

channel and a nonselective cation channel.

Radioimmunoassays of cAMP in preparations containing the outer

dendritic segments of the receptor cells, where odor reception and

transduction occur, revealed only a low level of adenylate cyclase

activity, and no significant changes were evoked by odorants.

Intracellular recordings demonstrated that forskolin or cAMP analogs

mixed with a phosphodiesterase inhibitor and applied to the outer

dendritic segments did not evoke responses directly or cause the

cell's odor response to adapt. These results do not support the

hypothesis that cAMP mediates transduction in lobster olfactory

receptor cells.

A mixture of L-arginine, L-proline, and L-cysteine evoked

hyperpolarizing receptor potentials in both P. argus and the

American lobster, Homarus americanus, providing an explanation for

noncompetitive mixture suppression in odor responses. All cells

hyperpolarized by this mixture were also depolarized by other

odorants, indicating that multiple transduction mechanisms may exist

in the same cell.

Histamine directly gated a chloride channel in receptor cell

somata. This channel required external histamine for gating, but

not GTP-binding proteins, metabolic energy sources, or Ca**. The

conductance of the channel was 66 pS in P. argus and 44 pS in H.

americanus. The pharmacology and dose-dependency of this channel

indicate that it mediates the suppression of excitability of the

lobster olfactory receptor cell caused by histamine.



Sensory systems provide an animal with the environmental

information necessary to make adaptive behavioral responses. An

understanding of how sensory systems perform this function begins

with the role of the receptor cell. Receptor cells must receive

stimuli, transduce stimulus energy into generator potentials or

action potentials, and transmit the electrical signals to the CNS.

They may also participate in the filtering of excessive or

adaptively insignificant signals. The molecular mechanisms

underlying these processes must be understood because they determine

which stimuli a sensory system is capable of detecting and how the

stimulus is encoded into the neural signals which are the internal

representation of an organism's environment. Receptor cells in the

vertebrate visual system are the most thoroughly investigated

examples (reviewed by Stryer, 1986). In photoreceptors, light

activates rhodopsin, which activates transducin, which activates a

phosphodiesterase. The phosphodiesterase catalyzes the hydrolysis

of cGMP, which decreases the cGMP-activated inward current, and

consequently decreases synaptic activity. This decrease in synaptic

activity is then the signal which is processed by the retina and

passed on to higher neural levels.

In comparison to other receptor cells, especially

photoreceptors, the olfactory receptor cell is poorly understood.

The non-spatial and physically discrete nature of olfactory stimuli

increases the complexity of determining stimulus quality. Unlike

vertebrate taste, where recognition of four basic qualities

simplifies categorization of responses, and where recent evidence

suggests that stimulus quality could be determined by distinct

transduction mechanisms (reviewed by Kinnamon, 1988), olfactory

responses appear to contain infinite variety. The unique features

of olfaction provide an opportunity to investigate mechanisms of

sensory coding, and ultimately perception, which may contrast with

known mechanisms in unforeseen ways, potentially contributing to a

more broadly based understanding of sensory coding.

Olfactory receptor cells detect and encode chemical signals.

They provide an organism with information about its chemical

environment, especially about low concentrations of odorants, often

emanating from distant sources. The activation of olfactory

receptor cells by odorants is a complex phenomenon which can be

divided into 5 processes: the perireceptor events, the binding of

odorants to receptors, olfactory transduction, the spread of odor-

evoked currents, and the effects of these currents on the generation

of action potentials. These processes have critical,

interdependent roles in the activation of the receptor cell.

Perireceptor events involve proteins which facilitate the

access of odorants to receptor sites, as well as proteins which help

to clear the odorants away from the receptor sites, as occurs in

synaptic clefts. Odorant-binding proteins, which seem to carry

odorants to, or concentrate odorants near, the receptor sites have

been characterized in insects (Vogt, 1987) and mammals (Bignetti et

al., 1985; Pevsner et al., 1985, 1988). Clearance systems such as

degradative enzymes and uptake systems have been observed in both

insect (Vogt et al., 1985) and mammalian olfactory organs (Dahl,

1988). In lobster olfactory organs, ectonucleotidases which degrade

adenine nucleotides, and uptake systems for adenosine and taurine,

have been described (Trapido-Rosenthal et al., 1987; Gleeson et al.,


Once odorants reach the olfactory receptor cells, they are

believed to bind to receptor proteins (reviewed by Lancet, 1988)

which presumably reside in the membranes of the dendrites (Adamek et

al., 1984). Binding studies using volatile odorants have proven

difficult because these lipophilic odorants can bind nonspecifically

to hydrophobic surfaces (reviewed by Lancet, 1988). But the binding

of amino acids to chemosensory membranes of fish has provided

evidence for receptor proteins (Cagan and Zeiger, 1978; Brown and

Hara, 1981; Fesenko et al., 1983). Receptor proteins presumably

also exist in lobster olfactory receptor cells because the

identified odorants are amino acids and other small, water-soluble

compounds (Carr et al., 1984; Derby and Ache, 1984a).

Olfactory receptor proteins activated by the binding of

odorants must then generate an electrical response, a process called

olfactory transduction. Evidence for several mechanisms of

olfactory transduction now exist. Biochemical studies demonstrating

that volatile odorants stimulate adenylate cyclase in olfactory

cilia from frogs and rats (Pace et al., 1985; Sklar et al., 1986)

have received electrophysiological support. The perfusion of cAMP

into salamander olfactory receptor cells evokes an inward current

(Trotier and MacLeod, 1987), and both cAMP and cGMP activate a

nonselective cation channel in excised patches from frog olfactory

cilia (Nakamura and Gold, 1987). Application of 8-bromo-cAMP,

forskolin, or phosphodiesterase inhibitors to voltage-clamped sheets

of frog olfactory mucosa also gives results which are consistent

with olfactory transduction by cAMP (Desimone et al., 1988). In

contrast, amino acids stimulate phosphoinositide turnover in catfish

olfactory cilia (Huque and Bruch, 1986), and a novel odorant-gated

channel has been reported in recordings of ion channels from frog

olfactory cilia reconstituted in planar bilayers (Labarca et al.,

1988). Although this evidence is impressive, especially for cyclic

nucleotides, proof that any of these mechanisms is necessary for

olfactory transduction is still lacking.

Once transduction has generated an ionic current, the current

must spread to interact with the voltage-dependent ion channels in

the spike generating zone. Because olfactory receptor cells are

bipolar neurons, odor-evoked currents must spread from the most

distal sections of the dendrites to the axons where action

potentials appear to be generated (Masukawa et al, 1985; Trotier,

1986; Schmiedel-Jakob, et al., 1989). Current spread is either

passive, or active mechanisms (i.e., voltage-dependent channels)

must participate. Present evidence from amphibian olfactory

receptor cells, including high membrane resistance and the apparent

absence of voltage-dependent currents whose properties would allow

them to enhance the spread of current (Firestein and Werblin, 1987;

Trotier, 1986; Masukawa et al., 1985), is consistent with passive


When the odor-evoked current reaches the spike generating

zone, the resulting change in membrane potential will change the

spike frequency. The pattern and frequency of these spikes

presumably encode information about the quality and quantity of the

odorants present (reviewed by Lancet, 1986). The characteristics of

the spike trains generated in response to excitatory odorants, such

as frequency adaptation and maximum and minimum spike frequencies,

are functions of the voltage-dependent currents of the receptor

cell and the underlying receptor potential. At present, our

knowledge of voltage-dependent currents and the ion channels which

underlie them in olfactory receptor cells is sketchy at best. Some

differences in channels and currents in olfactory receptor cells

occur between amphibian species (Trotier, 1986; Firestein and

Werblin, 1987; Suzuki, 1987), and greater differences occur between

amphibian and mouse olfactory receptor cells (Maue and Dionne,

1987). This suggests that a number of possible solutions exist for

the problem of regulating the frequency of spike trains.

These activation processes may not always proceed without

intervention from other parts of an animal's nervous system,

however. Evidence that receptor cell excitability may be regulated

by neuroactive substances suggests that the CNS may feed back onto

olfactory receptor cells. In frogs, substance P, perhaps released

from the trigeminal system, affects the electrical activity of the

olfactory receptor cells (Bouvet et al., 1987). In mosquitoes,

blood-borne factors cause decreased sensitivity of extracellularly

recorded responses to odorants (Davis and Takahashi, 1980). In the

olfactory receptor cells of lobsters, histamine suppresses both

spontaneous and odor-evoked spiking and decreases the amplitude of

receptor potentials (Bayer et al., submitted). The mechanisms

mediating these regulatory phenomena were unknown, until the

evidence presented in Chapter 6 demonstrated that a ligand-gated

channel mediates the action of histamine in lobster olfactory

receptor cells.

The Olfactory Organ of the Spiny Lobster

An attractive system for studying the activation of olfactory

receptor cells is the olfactory organ of the spiny lobster,

Panulirus argus. The odorants which have been identified are amino

acids, nucleotides, and other small water-soluble molecules (Carr et

al., 1984; Derby and Ache, 1984a). Some of these molecules are

neurotransmitters and neuromodulators, providing interesting

comparisons with their known mechanisms of action at internal sites

of chemoreception, such as synapses (Carr et al., 1987; 1989). The

olfactory receptor cells of the spiny lobster have long been the

focus of electrophysiological studies (reviewed by Ache and Derby,

1985). Extracellular recording studies have made significant

contributions to our understanding of olfactory coding phenomena,

such as mixture suppression (Derby and Ache, 1984a; Derby et al.,

1984; Gleeson and Ache, 1985; Ache et al., 1988) and discrimination

of complex mixtures of odorants (Derby and Ache, 1984b; Girardot and

Derby, 1988). The pioneering study of P. argus by Anderson and Ache

(1985), was the first in which patch-clamp methods were used to

obtain reliable intracellular recordings of odor responses from an

olfactory receptor cell. With methods and preparations for

intracellular and patch-clamp recording now available, the ability

to study the electrical properties and mechanisms of activation of

the olfactory receptor cells is enhanced.

A significant advantage of the olfactory organ of P. argus is

its morphology. The primary olfactory organ consists of the

aesthetasc sensilla located on the lateral filaments of the paired

first antennae, or antennules (Fig. 1-1). The ultrastructure of

these aesthetascs was recently described by GrUnert and Ache

(1988). An aesthetasc consists of a cluster of about 320 bipolar

sensory neurons, each of which sends a dendrite into the 800 pm long

cuticular hairs (Fig. 1-2). There may be up to 2000 aesthetascs and

600,000 olfactory receptor cells per olfactory organ. The dendrites

of the receptor cells branch repeatedly into outer dendritic

segments which extend into the distal 700 pLm of the aesthetasc hair.

The only other cellular material inside the hairs is the processes

of the auxiliary cells which ensheathe the inner dendritic segments

in the proximal 100 pLm of the hair. The inner dendritic segments

are 300 400 pLm in length and are about 0.5 pum in diameter. The

outer dendritic segments are 600 to 700 pm in length and 0.05 to

0.08 Im in diameter. In the base of the hair, septate junctions and

desmosomes between auxiliary cells, and desmosomes between auxiliary

cells and inner dendritic segments, appear to prevent the free

exchange of solutions between the hair lumen and the hemolymph space

in the antennular lumen. Auxiliary cells also ensheathe the

neuronal clusters in the antennular lumen. The axons run

uninterrupted from the neuronal cell bodies to the olfactory

neuropil of the CNS. This morphology provides several advantages.

Dissection is made easy by the exposed location of the organ. By

simply cutting off the hairs one can obtain an almost pure

preparation of olfactory membranes presumably containing all the

biochemical machinery necessary for the initiation of odor

responses. The physical separation of the solutions bathing the

outer dendritic segments and the soma allows selective application

of different solutions to these parts of the cells.

Statement of Purpose

The first objective of this project was to establish a

foundation of information about the ion channels and ionic currents

in the olfactory receptor cell of the spiny lobster. Description of

these channels and currents helped me to formulate several

hypotheses about mechanisms involved in the spread of odor-evoked

currents, olfactory transduction, inhibitory odor responses, and the

regulation of excitability of the receptor cell. From these

hypotheses I made four predictions.

The second objective was to test the predictions, which were as

follows. (1) A weakly voltage-dependent ion channel carrying inward

current enhances the spread of odor-evoked inward currents. (2)


Odorants stimulate adenylate cyclase to produce cAMP as a mechanism

of olfactory transduction. (3) Hyperpolarizing receptor potentials

exist. (4) A ligand-gated channel mediates the suppression of

olfactory receptor cell excitability by histamine.

Fig. 1-1. A. A drawing of the spiny lobster showing the location of
the aesthetasc sensilla on the lateral filaments of the antennule.
B. Four annuli of the antennule. The shaved aesthetasc hairs on the
first antennular section show the parallel arrangement of the bases
of the double row of hairs before they bend into the zig-zag
relationship formed by their tips. Reproduced from GrUnert and Ache
(1988), with permission.

Fig. 1-2. The aesthetasc sensillum of Panulirus argus. A. A
diagram of an aesthetasc sensillum in longitudinal section. The
proportions of the structures inside the hair are not drawn to
scale. B. A drawing of a longitudinal section of the region of the
hair where the dendrites branch. Axon (a); basal body (bb);
bulbous region (br); cuticle (c); ciliary segment (cs); endocuticle
(en); exo- and epicuticle (ex); inner auxiliary cell (iac); inner
dendritic segment (ids); intersegmental junction (ij); lymph space
(1); outer auxiliary cell layer, (oac); outer dendritic segment
(ods); pore-like structure (p); ciliary rootlet (r); sensory cell
soma (sc). Reproduced from GrUnert and Ache (1988), with






Olfactory receptor cells in diverse organisms are primary

bipolar sensory neurons that serve both to transduce the stimulus

and generate action potentials that propagate to the CNS.

Transduction is thought to occur in the terminal arbor of the

dendrite, which is exposed to the odor environment (reviewed by

Getchell and Getchell, 1987). Whole-cell patch clamp recordings in

amphibians (Firestein and Werblin, 1987; Trotier, 1986) and

crustaceans (Schmiedel-Jakob et al., 1989) indicate that olfactory

receptor cells have a high membrane resistance, permitting odor-

evoked receptor potentials to invade the soma and generate action

potentials in or near the axon. These intracellular findings

support similar interpretations of extracellular data from olfactory

receptor cells in vertebrates (Getchell, 1973). Extracellular data

from insect olfactory receptor cells are as yet inconclusive as to

the site of origin of action potentials (reviewed by Kaissling,

1986; De Kramer and Hemberger, 1987). Because the soma is

critically interposed between the transduction site and the output

of the cell, its voltage-dependent ionic currents, as well as its

passive membrane properties, play a critical role in determining

spike output by providing voltage-activated responses such as

rectification or afterhyperpolarization. Furthermore, because the

soma is usually the recording site and the injection site of drugs,

the contribution of the soma to experimental manipulations designed

to affect the dendrites must be considered. For these reasons, an

understanding of the somatic ion channels and currents is important.

Analyses of the ionic currents in olfactory receptor cells

have previously been confined to amphibians (Firestein and Werblin,

1987; Trotier, 1986; Suzuki, 1987). Collectively, amphibian

olfactory receptor cells possess a transient Na' current, a Ca**

current, a delayed rectifier current, a transient outward current (A

current), a Ca*-activated K' current, and perhaps an anion current.

As these currents were recorded from cells with intact dendrites,

the role of channels in the somatic membrane is unclear. In order

to properly interpret whole-cell recordings of odor-evoked currents,

to better understand the spread of these currents within the cells,

and to further the comparative perspective of olfactory receptor

cell function, I have investigated the ionic currents of somata

isolated from olfactory receptor cells of the lobster, P. argus.

This chapter describes the properties of four voltage-dependent

currents from olfactory receptor cells. These currents are a TTX-

sensitive Na* current, a Ca-* current, a delayed rectifier K-

current, and a Ca-'-activated K* current. How these currents could

produce the voltage-dependent responses of the olfactory receptor

cell -- action potentials, membrane repolarization, outward

rectification, and repetitive discharge of action potentials -- is



Preparation of isolated olfactory receptor cell somata

The lateral filament of an antennule (the olfactory organ) was

excised from specimens of P. argus, and cut into 1 mm hemisections

(Anderson and Ache, 1985). This procedure removes the majority of

the axon from each soma. Hemisections were agitated with L-

cysteine-activated papain (.25 mg/ml; Sigma, type IV) for 25 min and

then with trypsin (2.5 5 mg/ml in Ca"1 free saline; Sigma, type

IX) for another 25 min. Somata were isolated and stripped of their

dendrites by drawing the somata clusters into a glass pipette with a

200 pm opening and forcing them back out into a 35 ml plastic

culture dish, where they were allowed to settle and stick to the

bottom. Olfactory receptor cell somata were identified by their

spherical shape and 10 20 pLm diameter. An ultrastructural study

found that virtually all cells of this size and shape in the

olfactory organ are olfactory receptor cells (GrUnert and Ache,


Voltage-clamp recording

Patch clamp pipettes were made from borosilicate glass tubing

(Boralex, Rochester Scientific, Rochester, NY) pulled to a tip

diameter of just over 1 pm (Bubble numbers of 4.6 to 5.4, Mittman et

al., 1987). Sylgard 184 (Dow Corning, Midland, MI) was applied to

the neck of the pipette to reduce capacitance. After forming a

gigaohm seal and compensating the capacitance of the electrode, the

whole-cell recording configuration was achieved using suction to

break through the cell membrane and the series resistance error was

compensated qualitatively. A commercial voltage-clamp and current

amplification circuit with a 1 gigaohm head stage (Dagan Instr.

model 8900, Minneapolis, MN) was used to record macroscopic currents

from the somata. An IBM-XT computer with a D/A, A/D converter and

accompanying software (pClamp, Axon Instr., Burlingame, CA) was used

to apply voltage-step protocols and to store digitized data for

later analysis. Data were low-pass filtered at a corner frequency

of 10 kHz. Leak currents were recorded from each soma by applying

half amplitude voltage step protocols at hyperpolarized potentials

where no active currents were elicited. Leak correction was then

performed by multiplying these leak currents by + 2 as necessary,

and subtracting them from the active current records. Recordings

were made at room temperature, which varied from 22 to 240C.


The composition of lobster saline was (mM) 480 NaC1, 13 KC1,

13 CaClz, 10 MgCl2, 1.7 glucose, 20 HEPES, pH 7.4. The composition

of the normal patch solution used was (mM) 200 KC1, 2 MgCl2, 11

EGTA, 1 CaClz, 570 glucose, 10 HEPES, 34 NaOH, pH 7.4. The

compositions of other solutions used are described in the figure

legends. Inorganic salts were purchased from Fisher Scientific

(Fair Lawn, NJ) and HEPES and EGTA were purchased from Research

Organics (Cleveland, OH). All other chemicals were purchased from

Sigma Chemical Co. (St. Louis, MO).


Total membrane currents

In response to depolarizing voltage steps, isolated somata

bathed in lobster saline and perfused with normal patch solution

showed an inward current, followed by a much larger outward current

(Fig. 2-1). Hyperpolarizing voltage steps down to -120 mV, from

holding potentials of -35 to -70 mV, failed to activate any currents

in these cells. The inward current became apparent at slightly more

negative potentials (-40 to -30 mV) than did the outward current

(-30 to -20 mV). Specific blocking agents and impermeable ions were

then used to isolate four component currents from the total membrane

currents of these somata. In no instance did a cell fail to

exhibit the appropriate current for the recording conditions used.

Na' current

Replacing the K- in the patch solution with Cs* and adding 4

mM Co-* or Cd*- to the bath allowed isolation of a fast, transient

inward current (N = 24 cells) that peaked within 1 ms of the

initiation of the voltage step (Fig. 2-2). This current was absent

in the presence of submicromolar concentrations of TTX (N = 111

cells) and saxitoxin (N = 2 cells), identifying this as a Na-

current. The current activated at -38 to -30 mV and reached maximum

amplitude at 0 to 10 mV. The rapid onset of the current at -6 mV

was described by a single exponential with a time constant of 0.07

0.03 ms (N = 7 cells). The current decayed as a single exponential

function with a time constant of 0.29 0.08 ms (N 13 cells).

Large Na* currents were not correlated with cell size as measured by

Fig. 2-1. Total membrane currents of an isolated soma. Above:
Macroscopic currents evoked by depolarizing voltage steps (inset)
showing a small, biphasic inward current and a large outward
current. Leak currents were not subtracted from these records.
Below: The I-V relationships of the inward current (circles) and the
outward current (triangles) from this cell. Pipette: normal patch
solution. Bath: saline.

O mV
0 -40mV
-70mV- -

-30 -10 10 30








Fig. 2-2. The Na- current. Above: Transient inward currents
elicited by depolarizing voltage steps (inset). Leak currents were
not subtracted from these records. Below: The I-V relationship of
this current. Activation began between -40 and -30 mV. The peak
current occurred at 0 to 10 mV. Pipette (mM): 200 CsCl, 5 4-
aminopyridine, 11 EGTA, 1 CaC12, 565 glucose, 10 HEPES, pH 7.4.
Bath: saline, plus (mM); 4 CoC12, 4 CdCz1, 10 TEA.


0 /0
\ .-0









linear regression of current amplitude at -6 mV versus cell

capacitance (correlation coefficient = 0.30, N = 13 cells), but

somata with a neurite, presumably an axon, had larger Na currents,

suggesting that axons contain higher densities of Na' channels.

Steady state inactivation was half maximal at -48 mV (Fig. 2-3A; N =

7 cells). Inactivation was evoked by prepulses to a series of

increasing potentials for 50 ms before the membrane potential was

brought to 0 mV to elicit the Na' current. Recovery from

inactivation followed a single exponential function with a mean

time to half maximal activation of 1.7 0.4 ms (N = 3 cells) at

room temperature (Fig. 2-3B). Recovery from inactivation was

measured using a two pulse protocol in which a series of paired

pulses to 0 mV from a holding potential of -70 mV were separated by

an increasing interval of time.

Ca"* current

A sustained inward current was isolated when Cs- replaced K- in

the patch solution and 0.1 pM TTX was added to the bath (Fig. 2-4; N

- 44 cells). This current, which washed out within minutes after

breakthrough into the whole-cell configuration, was carried by

either Ca*-, Ba-1, or Sr"' and was absent when either Co"* or Cd+",

or both, was added to the bath at a concentration of 4 to 10 mM (N -

65 cells). These observations identified this current as a Ca-"

current. Like the Na current, this current activated between -40

and -30 mV. When carried by Ca-* (N 4 cells), the current decayed

more rapidly (Fig. 2-5), suggesting that a Ca-dependent process is

involved in the inactivation of this current. The washout of this

current was not suppressed by 2 mM ATP and 0.2 mM GTP (N 5 cells).

Fig. 2-3. Properties of inactivation of the Na- current. A. Steady
state inactivation (ho) of the Na- current elicited by 50 ms
prepulses to the potentials depicted on the abscissa. Shown are the
means and standard deviations from four cells. Small standard
deviations fall within circles marking the data points. B.
Recovery from inactivation evoked by a paired pulse protocol. The
plot denotes the amplitude of the Na- current evoked by the second
pulse relative to the first versus the interval between the two
pulses. The data points were fit by a single exponential curve
with a time constant of 1.7 ms. Solutions as in Fig. 2-2.




3 5 8 10







0.0 -











_ 1 1 1 1


Fig. 2-4. The Ca-- current. Above: Sustained inward currents
carried by Sr"* elicited by depolarizing voltage steps (inset).
Leak currents were subtracted from these records. Below: The I-V
relationship of the current from the same cell. Activation began
at about -40 mV. Peak current occurred at about -10 mV. Pipette
(mM): 150 CsCl, 50 TEA, 5 4-aminopyridine, 2 MgClz, 1 CaCl2, 11
EGTA, 2 ATP, 0.2 GTP, 565 glucose, 10 HEPES, pH 7.4. Bath (mM):
480 NaCl, 13 KC1, 20 SrC12, 3 MgC12, 5 TEA, 20 HEPES, pH 7.4, and
0.1 ILM TTX.

o_ *

Z -200--

3 -400-- a ,*

-60 -30 0 30


Fig. 2-5. Currents through Ca-* channels evoked by depolarizing
voltage steps to 0 my (inset). Leak currents were subtracted from
these records. Currents decayed more rapidly when carried by Ca--
(upper trace) than when carried by Sr-- (lower trace). Currents
shown are from two different somata, normalized to the same
amplitude and superimposed. Solutions as in Fig. 2-4 except that
Ca"' replaced Sr"* in the saline bathing one cell.

Ca*-activated K' current

With 5 mM TEA and 0.1 p.M TTX in the bath, a very large outward

current was apparent (Fig. 2-6; N = 12 cells). Like the Call

current, this current disappeared within minutes after breakthrough

into the whole-cell configuration. This washout was not suppressed

by 1 mM dibutyryl cAMP (N = 3), which prevents washout of Call

currents in some neurons (Chad and Eckert, 1986). The current was

absent if 4 mM Co- or Cd-- was added to the bath (N 55 cells) or

if internal K* was replaced by Cs' or Na- (N 64 cells). These

observations identify this as a Ca+-activated K- current. Applying

20 mM TEA internally (N 2 cells) or 5 mM TEA externally (N = 6

cells) failed to block the current. The Cal'-activated K- current

decayed to varying degrees during voltage steps, apparently due to

the inactivation of the Cal current, as the single channel

underlying the Ca-activated K* current shows no inactivation

(Chapter 3).

Delayed rectifier K' current

A delayed outward current was isolated by adding 0.1 PM TTX

and 4 mM Co* or Cd-*, or both, to the bath (Fig. 2-7; N 55

cells). This current activated with a delay and inactivated slowly.

The current was blocked by 5 mM external TEA (N 6 cells) and was

absent when internal K* was replaced by Cs* or Na- (N 60 cells).

Collectively, these features identify this current as a delayed

rectifier type of K current. Inactivation of this current was

voltage dependent (Fig. 2-8). The inactivation elicited by 250 mS

prepulses, which were too brief to achieve steady-state

inactivation, was half-maximal at -41 mV (N 4).

Fig. 2-6. The Ca+--activated K- current. Above: The macroscopic
currents evoked by depolarizing voltage steps (inset) immediately
after breakthrough into the whole cell recording configuration.
The Ca" current providing the internal Ca-" for activation can be
seen as the small inward current prior to the outward current.
Leak currents were subtracted from these records. Below: The I-V
relationship of the peak Ca---activated K+ current from the same
cell immediately after breakthrough (circles) and 5 minutes later
(triangles) illustrates the rapid washout of the Ca+- current.
Pipette: normal patch solution plus 20 mM TEA. Bath: saline plus 4
mM TEA and 0.1 p.M TTX.

r0: 2.0

-50 -30 -10 10 30 50


Fig. 2-7. The delayed rectifier K' current. Above: Macroscopic
currents evoked by depolarizing voltage steps (inset) under
conditions allowing isolation of the delayed rectifier current.
Below: The I-V relationship of the peak current from the same cell.
The current began to activate between -30 and -20 mV. Pipette:
normal patch solution plus 2 mM ATP and 0.2 mM GTP. Bath: saline
plus 4 mM CoC12, 4 mM CdClz and 0.1 pM TTX.

1.8j I I I









0.6 !


Fig. 2-8. The voltage dependence of inactivation of the delayed
rectifier current under conditions where steady-state inactivation
had not been reached. The plot shows the percent of maximum current
evoked at 40 mV versus the prepulse potential. The mean and
standard deviation of 4 cells are plotted. Prior to stepping the
holding potential to 40 mV to evoke the delayed rectifier current,
250 ms prepulses were applied in positive steps of 7.5 mV beginning
at -110 mV. Between voltage step protocols the membrane was held at
-70 mV. Solutions as in Fig. 2-6.

z *

25 1

0 I- ,

-i20 -80 -40


This current was also modulated by at least one mechanism

involving a GTP-binding protein. Perfusing somata with normal patch

solution containing 20 mM NaF and 10 iM AlCls, a mixture that

generates AlF4- which nonselectively activates GTP-binding proteins

(Blackmore et al., 1985; Sternweis and Gilman, 1982), reduced the

delayed rectifier over time (Fig. 2-9A vs. B). Analysis of

variance showed that the reduction of the current in 5 somata

perfused with this mixture was significantly greater than that of 7

somata perfused with normal patch solution (p < 0.05). Perfusing

somata with normal patch solution containing 1 mM GTP-r-S, another

GTP-binding protein activator, caused similar reductions in the

current over time (Fig. 2-9C). Analysis of variance showed that

the reduction of current in 5 somata perfused with GTP-r-S was

significantly greater than that of the 7 somata perfused with normal

patch solution (p < 0.05).

The delayed rectifier current was also blocked by forskolin.

External application of 10 JLM forskolin (N = 5 cells) caused a

voltage-dependent increase in the rate of inactivation and decreased

the amplitude of the delayed rectifier (Fig. 2-10). This particular

effect of forskolin, however, has recently been shown to be a direct

block of K* channels and not mediated by forskolin's ability to

stimulate adenylate cyclase (Watanabe and Gola, 1987; Hoshi et al.,



Depolarizing receptor potentials evoked by odorants in lobster

olfactory receptor cells can be larger than 40 mV and have plateau

Fig. 2-9. Effect of agents which activate GTP-binding proteins on
the delayed rectifier current. A. The I-V relationships of the
delayed rectifier current from an untreated cell showed little
reduction over time. Plotted are current amplitudes immediately
after breakthrough (circles), two minutes later (triangles), and 6
minutes later (squares). Solutions as in Fig. 2-6. B. The I-V
relationships from a cell perfused with AlF4- shows reduction of
the delayed rectifier current over time. Plotted are current
amplitudes from immediately after membrane breakthrough (circles), 3
minutes later (triangles), and 6 minutes later (squares). Solutions
as in Fig. 6 except that 20 mM NaF and 0.1 LM AlC13 were added to
the patch solution. C. The I-V relationships from a cell perfused
with GTP-F-S also shows reduction of the delayed rectifier current
over time. Plotted are current amplitudes immediately after
breakthrough (circles), two minutes later (triangles), and 5 minutes
later (squares). Solutions as in Fig. 2-6, except that 0.5 mM GTP-
r-S was added to the patch solution.



Z 800

D 400




-30 -10 10 30


v 400


Fig. 2-10. The effect of forskolin upon the delayed rectifier
current. Above: An untreated cell showed a normal family of
delayed rectifier currents. Below: In the presence of 10 pM
forskolin, the delayed rectifier current in a neighboring cell
showed voltage-dependent inactivation. Solutions as in Fig. 2-6
except that forskolin was added to the bath between recording from
the two cells.


I 30mV
400pA '-40mV
3ms -70mV- OV


__ = -40mV
3ms -70mV-

potentials lasting at least a few hundred milliseconds (Anderson

and Ache, 1985; Schmiedel-Jakob et. al., 1989). My results support

the hypothesis that these receptor potentials spread passively from

the dendrite and generate Na* action potentials in the axon. I

found no evidence of a voltage-dependent inward current active at

potentials more negative than the Na* current, or of an ongoing

outward current decreased by potentials between rest and threshold,

either of which could increase the effective space constant of the

cell (Yoshii et al., 1988). This suggests that the spread of

receptor potentials, at least through the soma, is passive. Unless

the dendrite has voltage-dependent currents not found in the soma,

one must assume that the passive membrane properties of the cell

must account for the spread of current. My observation that the

size of the Na* current correlated not with the size of the soma,

but with the presence of a remnant of neurite, suggests that most

Na* channels are located on the neurite. Assuming that the

neurites were axons would be consistent with intracellular

recordings from intact cells indicating that the axon is the site

of generation of action potentials in lobster olfactory receptor

cells (Schmiedel-Jakob et al., 1989). Furthermore, the relatively

negative potentials at which the Na* current inactivates suggests

that any Na* channels present in the soma and dendrite would rapidly

inactivate during a receptor potential. Trotier (1986) similarly

concluded that the axon contains most of the Na* current based on

its infrequent occurrence in isolated salamander olfactory receptor


Because 1 mM external TEA broadens action potentials and

reduces outward rectification in these cells (Schmiedel-Jakob et

al., 1989), the delayed rectifier K- current may participate in

both repolarization of the action potential and in the outward

rectification of the cell. The Ca-*-activated K- current could also

be involved in both these functions as well because it activates

almost as rapidly as a similar current involved in repolarization of

the action potential in bullfrog sympathetic neurons (MacDermott and

Weight, 1982) and because it takes more than 100 ms to decay.

Predictions about functions of the Ca-activated K' current have

been difficult to test, however, due to the rapid washout of the

Call currents in the whole-cell recording configuration.

Outward rectification produced by the K* currents may act to

expand the range of odor-evoked inward currents which can be

encoded into changes in spike frequency. Given that the interval

of potential from threshold to saturation of spike frequency is

fixed in any one cell, from Ohm's law, the amount of inward current

capable of causing potentials that fall within this interval would

increase as the K* currents decrease the membrane resistance. That

this occurs is demonstrated by TEA block of the delayed rectifier

current: the injection of 100 pA of current into an intact lobster

olfactory receptor cell caused a potential change of 23 mV before,

and 47 mV after, TEA application (Schmiedel-Jakob et al., 1989).

Unlike amphibian olfactory receptor cells (Firestein and

Werblin, 1987; Trotier, 1986; Suzuki, 1987), the somata of lobster

olfactory receptor cells lack an A current or other transient

outward current. In many neurons, an A current appears to be

necessary to allow repetitive spiking over a wide range of

frequencies (Connor, 1976). How the lobster olfactory receptor

cell accomplishes repetitive spiking over a wide range of

frequencies (Schmitt and Ache, 1979) in the absence of an A current

is unclear. Perhaps one or both of the K* currents subserves this

function by slowing the return to threshold following a spike. A

Ca*-activated K* current appears to permit repetitive spiking in

frog motoneurons (Barrett and Barrett, 1976). Mouse olfactory

receptor cells also appear to lack the K* channel underlying the A

current (Maue and Dionne, 1987), suggesting that multiple mechanisms

exist for controlling spike frequency in olfactory receptor cells.

The depression of the delayed rectifier current by forskolin

and by agents which activate GTP-binding proteins illustrates a

caveat for the somatic injection of agents intended to act upon

transduction mechanisms in neuritic processes. Care must be taken

to ensure that such agents have specific actions which occur only in

the desired location within the cell.

In conclusion, lobster olfactory receptor cells appear to

perform the same functions as their amphibian counterparts, but

with a slightly smaller number of voltage-dependent currents. At

present, these results suggest that the Na- current, located mostly

in the axon, underlies action potentials in lobster olfactory

receptor cells, while the Ca** and K- currents shape the plateau of

the receptor potential, repolarize the membrane, and control

repetitive spiking. Recording of odor-evoked responses from intact


cells while selectively blocking each of the cell's ionic currents,

would test the proposed physiological roles of these four currents.




Olfactory receptor cells are primary sensory neurons that

transduce chemosensory information into graded electrical signals

and generate action potentials that propagate to the CNS. The

description of the ion channels which underlie these electrical

responses is a step toward understanding olfactory receptor cell

function. Because these are bipolar neurons, the ion channels of

the soma interact with odor-evoked membrane potential changes

spreading from transduction sites in the dendrites (reviewed by

Getchell and Getchell, 1987) to the axon where action potentials are

presumably generated (Getchell, 1973; Masukawa et al., 1985;

Trotier, 1986; Schmiedel-Jakob et al., 1989).

To date little information exists about the somatic ion

channels of olfactory receptor cells. Mouse olfactory receptor

cells have two Ca-activated K- channels, an inward rectifier, a

delayed rectifier, a Cl- channel, and a Call channel (Maue and

Dionne, 1987). K* channels have also been observed in salamander

olfactory receptor cells (Trotier, 1986). Because the morphology of

lobster and vertebrate olfactory receptor cells is similar (Grunert

and Ache, 1988), a comparison of ion channel properties may indicate

electrical properties which are common to olfactory receptor cells

in general, and identify potential roles of the channels in the

function of lobster olfactory receptor cells.

In this chapter I characterize four ion channels found on the

soma of the lobster olfactory receptor cell and correlate their

properties with functions of the receptor cell. These channels are

a steady state Cl- channel, a Ca+-activated K' channel, a voltage-

activated K1 channel, and a weakly voltage-dependent cation channel.


Preparation of receptor cells

Specimens of the spiny lobster, P. argus, were collected in the

Florida Keys and maintained in running seawater for up to three

months. Olfactory receptor cells were prepared for patch-clamp

recording by treating hemicylinders of cuticle containing the cells

with enzymes as described in Chapter 2. Hemicylinders, or isolated

somata, were placed in a 35 mm culture dish for patch clamp

recording. Cells were viewed at 200X or 300X under brightfield or

modulation contrast optics.

Patch clamp recordings

Patch clamp pipettes were made from borosilicate glass pipettes

(H15/10/0181, Jencons Scientific Ltd., Leighton Buzzard, England)

pulled and fire-polished to tip diameters of less than 1 pm (bubble

numbers of 3 to 4; Mittman et al., 1987). Sylgard 184 (Dow Corning,

Midland, MI) was applied to the electrode neck to reduce capacitance

when attempting to record channels rapidly activated by step changes

in holding potential. Single channel currents were amplified with a

patch clamp amplifier and a 10 gigaohm head stage (Dagan

Instruments, Minneapolis, MN). Records were low-pass filtered at 10

kHz for storage on video tape (Bezanilla, 1985) and at 0.5 to 5 kHz

for analysis, using an eight pole low-pass Bessel filter (Frequency

Devices, Inc., Haverhill, MA). An IBM XT with an A/D, D/A converter

and accompanying software (pClamp, Axon Instruments, Burlingame, CA)

was used both to apply voltage step protocols and to digitize and

analyze single channel records. Kinetic analysis was performed only

on records from patches containing one active channel. Open and

closed duration distributions were compiled as histograms and

exponential probability density functions were fitted to them by a

chi-square minimization method. For cell-attached patches, voltages

are reported as the voltage applied across the membrane, without

reference to the contribution of the cell's membrane potential (not

measured), following the convention of inside as negative. Values

are reported as means + standard deviations.


The compositions of patch solutions and salines are listed in

Table 3-1. HEPES and EGTA were obtained from Research Organics,

Inc. (Cleveland, OH) and inorganic salts from Fisher Scientific

(Fair Lawn, NJ). All other chemicals were obtained from Sigma

Chemical Co. (St. Louis, MO). Calcium concentrations in EGTA

solutions were calculated according to Hagiwara (1983). Solutions

were buffered to pH 7.4 with HEPES.

Solution changes were made by moving inside-out patches through

a slot cut into 2 concentric Delrin rings, one of which was fixed in

TABLE 3-1. Composition of solutions used. in mM

Saline Low Ca-* Ba-- KC1 NaC1 Low Cl Na

divalent saline saline saline saline saline patch


NaCI 460 480 420 420 480 350 190

KC1 13 13 480

CaC12 13 .01 60 .01 .01 .01 1

MgC12 10

BaC12 60


HEPES* 10 20 10 10 20 20 10 20

glucose 1.7 1.7 1.7 598

NaOAc 130

* Either NaOH or KOH was used to adjust the pH of solutions in order
to allow preparation of pure Na- and K- salines.

the bath. The outer ring was then turned about the fixed inner

ring to isolate a 300 Li chamber from the remainder of the bath. A

push-pull syringe system allowed changing the solution in the

chamber independently of that in the remainder of the bath.

Alternatively, the patch was moved into a flow from a glass pipette

(100 pm tip diameter) connected by a switching valve to six solution



A voltage-dependent cation channel

Conductance and effects of monovalent ions. The most

frequently observed channel in inside-out patches was a

nonselective cation channel. This channel was never clearly

observed in cell-attached patches. It had a linear I-V relationship

with low divalent saline on its cytoplasmic face and Na- patch

solution on its extracellular face (Fig. 3-1A, B). The mean slope

conductance was 320 21 pS (N = 19). Replacing Na- with K- in the

solution on the channel's extracellular face or with Cs- on its

cytoplasmic face caused the reversal potential to shift to values

intermediate between the Nernst potentials of Na and K- or Cs*.

Permeability ratios for K- relative to Na- were 0.66 and 0.74 for

the two channels tested, and for Cs- relative to Na- were 0.33 and

0.35 for two other channels. Substituting glutamate or gluconate

for Cl- on the cytoplasmic face (N 3) failed to shift the reversal


Replacing the 480 mM Na- in the intracellular solution with Cs-

or K- caused a gradual reduction in the frequency of channel

Fig. 3-1. A nonselective cation channel in an inside-out patch
bathed in Na- patch solution and low divalent saline. A.
Representative records of the channel under conditions where Na- was
the major charge carrier. Arrows indicate the current level of the
closed state. Corner frequency (f.) = 2 kHz. B. A plot of the I-V
of the channel depicted in A. The slope conductance was 323 pS.

20 .._.lJ I LL I

-60 I
s4 pA

-70 -35 0 35 70



openings and total lack of activity within a few minutes. Reducing

the Na- concentration to 20, 50, 100, or 200 mM by K- or Cs-

substitution caused the same effect. Adding back the 480 mM Na*

caused recovery of channel activity in only 3 of 8 patches. The

temporal correlation between these solution changes and changes in

channel activity was poor, suggesting that the effect was not a

simple Na4 activation of gating.

Single channel kinetics. Subsequent measurements of the

channel's properties were performed on inside-out patches with an

intracellular solution containing 460 or 480 mM Na-. With high

intracellular Na', the cation channel rarely became inactive, even

in patches which lasted longer than one hour. The probability of

being in the open state (Po) was increased by depolarization (Fig.

3-2A). Voltage dependence was analyzed by fitting these data to a

linearized form of the Boltzmann equation (Fig. 3-2B). Apparent

gating charges, calculated from the slopes of these lines, had a

mean value of 1.11 .06 (N 5). The average half-maximal

activation voltage was -37 7 mV. The increase in Po caused by

depolarization was due to both increased mean open times and

decreased mean closed times (Fig. 3-3). Most open dwell time

distributions were fit best by a single exponential (Fig. 3-4A)

whereas closed dwell time distributions usually were better fit by a

double exponential (Fig. 3-4B).

None of the properties of the cation channel were affected by

perfusing the intracellular side of the patch with salines

containing 10-10 M to 10-' M Ca4-, nor were they affected by 10 p!M

Fig. 3-2. The voltage dependency of the probability of the
nonselective cation channel being in the open state (Po). A. Plot
of the relationship of P, and voltage for one channel. The curve
was fit to the data (points) by nonlinear regression. B. Plot of
the Boltzmann linearization of the data from A. The line was fit
to the data by linear regression.




CL *



0 ..
0.0, --- ___

-120 -E0 -40 0 40
LT-E (mV)

Fig. 3-3. Plot of voltage dependence of the mean open (closed
circles) and closed times (open circles) of a nonselective cation
channel. Depolarization increased the Po by increasing the mean
open time and decreasing the mean closed time. The curves were fit
to the data by nonlinear regression.




-120 -80 -40 0 40



Fig. 3-4. Representative plots of the distributions of the open and
closed dwell times of a nonselective cation channel. Dwell times
were plotted as histograms and then exponential curves were fit to
the binned data by a chi-square minimization method. A. The
distribution of a record of 602 open dwell times at 20 mV fit by a
single exponential with a time constant of 13.9 ms. B. The
distribution of the closed dwell times from the same recording at 20
mV fit by a double exponential curve with time constants of 0.2 ms
and 2.0 ms.



A "i B


Z -\

Dwell Time (ms)

cGMP (N = 10) or 10 p.M cAMP (N = 8) applied to the same side.

External TTX (N = 1) and mouth suction or pressure applied to the

pipette also had no discernible effect. A potent odor mixture (a

filtered, saline extract of TetraMarin, TetraWerke, Melle, FRG)

applied in the patch pipette failed to cause the appearance of this

channel (or any other channel) in cell-attached patches or to alter

the properties of the cation channel in inside-out patches (N = 33).

Block by tetraethylammonium and divalent cations. Perfusion of

the intracellular side of the channel with salines containing 20 or

50 mM TEA caused a voltage dependent channel block characterized by

rapid flickering between the open and closed current levels (data

not shown). Block by TEA was not pursued further. Block by

internal divalent cations, however, provided clues about the pore

morphology and kinetic properties of this channel. Ba-", Ca"-,

Mn--, Co+*, and Mg--, at concentrations from 1 to 13 mM, all reduced

the open channel amplitude in a manner characteristic of very fast

blocking kinetics (Fig. 3-5A, B). The block by Mg--, the only

divalent cation likely to occur at millimolar intracellular

concentrations, was investigated in detail.

In fast block, Mg"- decreased the apparent open channel

amplitude in a concentration-dependent manner (Fig. 3-5A). The

dissociation constant of Mg"+ for the binding site, or sites,

responsible for fast block was obtained by analyzing the blocker

titration curve of Mg*- 'Fig. 3-6A) in which the ratio of the

blocked to the unblocked channel amplitude (iB/i) is the fraction of

time a channel is not occupied by the blocker (Moczydlowski, 1986).

Fig. 3-5. Representative records showing the blocking effect of
divalent cations applied to the intracellular side of inside-out
patches containing a nonselective cation channel. A. The
concentration dependence of block by Mg*". All records obtained
from the same patch containing 3 cation channels. B_. 10 mM or more
Mn--, Co--, and Ca'- appeared to reduce the open channel amplitude
below the limit of resolution. 10 mM Ba-" blocked the channel less
effectively than the other ions. These records were from several
different patches containing a single cation channel. All records
were obtained at -25 mV in symmetrical 480 mM NaCl and low pass
filtered at a corner frequency of 1.5 kHz. Arrows indicate the
current level of the closed state.

Mn" lOmM

Co" IOmM

Cao+ 13mM

Ba +,
OmM i3pA


Mgt IOmM


Mg +

At binding equilibrium between n blocker molecules and the open

channel (1), the equations describing the blocker titration curve

(2) and its Hill plot (3) are

nB + 0 nB-0 (1)

iB/i = (1 + [B]"/KB)-1 (2)

In(i/ia 1) = n ln[B] In Kn (3)

where [B] is the concentration of blocker, n is the Hill

coefficient, and K. is the dissociation constant of the blocker.

Hill coefficients and dissociation constants obtained from two

patches exposed to 1 to 10 mM Mg-* were 1.34 and 2.0 mM for one

patch and 1.61 and 6.4 mM for the other (Fig. 3-6B). Hill

coefficients of greater than one indicate that more than one ion is

involved in fast block.

Fast block by Mg-" was also voltage-dependent (N = 5), reaching

a maximum between -30 and -10 mV (Fig. 3-6C), possibly corresponding

to the Na- reversal potential (EN.) of -23 mV. This observation

suggested that Mg-- interacted with Na- in the permeation pathway

and that Mg- might even be slightly permeable. The voltage

dependence of the fast block was analyzed according to the Woodhull

(1973) model which assumes that voltage dependence arises because

the binding site for a charged blocking particle lies within the

membrane's electric field, for example, in the channel pore. The

data from membrane potentials below E.. were linearized and

estimates of Ka and of z6, the proportion of the membrane's electric

field sensed by a charged blocking particle at its binding site,

Fig. 3-6. Analysis of the Mg-* block of the nonselective cation
channel. A. A representative plot of the blocker titration curve
of one channel; the blocked single channel current normalized to the
unblocked current as a function of the concentration of Mg-1 at 0
mV. B. Plot of the fit of the Hill equation to the data in A. A
linear regression of these data yielded a Hill coefficient of 1.34
and a dissociation constant of 2.0 mM. C. Plot of the voltage
dependence of block of 3 mM Mg--; block is relieved to differing
extents by hyperpolarization and depolarization. D. A plot of
voltage versus the normalized reduction of the Po caused by
intracellular Mg-- shows that the effect of Mg"* on channel gating
is relatively independent of voltage. 1 mM Mg*- (circles), 3 mM Mg--
(squares), 5 mM Mg-" (triangles).



0 4 8 12
[Mg,"] (mM)
0.C0 *



-50 --.0 -,U 10

([Mg 2])

-50 -30

-10 10




C.5j t



were obtained from the following equation (Moczydlowski et al.,


ln[(i/ia) 1] = z8FV/RT + In([B]/Ka) (4)

Estimates of Ka averaged 39 16 p1M Mg-- (N = 5). Estimates of z6

averaged 2.6 0.2 (N = 5), indicative of multiple blocker binding


In addition to fast block, Mg*" also caused a concentration-

dependent reduction in Po, termed slow block, that was voltage

independent (Fig. 3-6D). The difference in voltage dependency

between fast and slow block suggests that distinct binding sites, or

sets of binding sites, mediate the two types of block. Slow block

was not limited to Mg--, as 10 mM Ba- applied internally reduced Po

from 0.84 to 0.25 at -20 mV.

Permeability to Ca"- and Ba-. The finding that holding

potentials more positive than EN. relieved the block by Mg raised

the possibility that divalent cations could permeate this channel.

To test this possibility, the internal faces of inside-out patches

containing cation channels were perfused with 60 mM divalent

cations. This manipulation, however, eliminated all channel

activity, possibly due to the slow block by divalent cations

described above. Subsequently, channel activity which may

correspond to the cation channel was easily found in inside-out

patches from cells bathed in low divalent saline when either Ca- or

Ba" saline was applied the extracellular side of inside-out

patches. Under these conditions, inward current events were

observed at potentials above EN. (-3 mV) and Ecx (-11 mV) where they

could only be carried by the divalent cation (Fig. 3-7A, B; Chesnoy-

Marchais, 1985). The properties of these divalent permeable

channels; no inactivation, weak voltage dependence, and presence in

inside-out, but not cell-attached patches, correlate with those of

the cation channel. The conductance of the channel in Ba" saline

was 80 pS, while values of 43 and 44 pS were obtained in Ca"* saline

(Fig. 3-7C). If these channels were the same cation channel found

in solutions with low divalent cation concentrations, the Po in Ba"*

saline was much reduced (e.g., Po = 0.39 at -5 mV versus > 0.9, cf

Fig. 3-2). The effect of Ca-" saline was even more dramatic in that

the Po increased to 0.1 at -10 mV, but failed to increase further

with stronger depolarization.

A cell-attached cation conductance. The nonselective cation

channel was never clearly apparent in cell-attached patches, even in

those which demonstrated cation channel activity after excision to

the inside-out configuration. With NaCl or KC1 solutions containing

micromolar divalent cations in the patch pipette, however, large

inward currents whose kinetics were too rapid to permit resolution

of the current amplitude occurred in cell-attached patches on cells

bathed in low divalent saline (Fig. 3-8). This cell-attached

channel activity was not observed with millimolar concentrations of

divalent cations in the pipette, and only rarely occurred in cells

bathed in salines containing millimolar concentrations of divalent

cations. Depolarization decreased the apparent amplitude and

increased the frequency of these current transients. When excised

into the inside-out configuration, 65Z of these patches contained

Fig. 3-7. Inward current events carried primarily by Ba-* and Ca--
through putative nonselective cation channels in inside-out patches.
A. Representative records showing inward current events in Ca"+
saline. The patch contained two channels, indicated by simultaneous
open events summing to twice the single channel current amplitude
(not shown). Arrows indicate the current level of the closed state.
f. 1 kHz. B. Representative records showing inward current
events in Ba-* saline. The patch contained three channels. f. = 1
kHz. C. Plot of the current-voltage relationships of the data
represented in A. and B. in external Ca-- saline (closed circles)
and in Ba-- saline (open circles). Note the inward rectification.
Curves were fit to the data by nonlinear regression.


SmV '- 4



ri r ''''


C 15ms
0.0- I i



-100 -75 -50 -25 0 25



Fig. 3-8. Representative records of a putative nonselective cation
channel in a cell-attached patch. Note the rapid transitions
between the open and closed states of the channel. This was the
record with the slowest kinetics observed. The pipette contained
NaC1 saline. Arrows indicate the current level of the closed state.
f= 2 kHz.

;''1 1 I :

( i i
40 mV 1



_MO 'Y 4pA

one or more of the familiar nonselective cation channels, but no

other channel activity. Less than 40Z of the silent cell-attached

patches showed cation channel activity after being excised.

Contingency table analysis of these proportions showed a significant

interaction of cell-attached and inside-out cation channel activity

(chi square 4.1, p < 0.05, 1 df, N 119). These observations

allow that the vastly different types of kinetic activity observed

could have arisen from a single species of cation channel responding

to the greatly different biochemical environments imposed by cell-

attached and inside-out recording configurations.

A Call-activated K channel

A large-conductance, Ca-*-activated K" channel was observed in

both cell-attached and inside-out patches. In symmetrical 480 mM

KC1, its I-V relationship (Fig. 3-9A) had a slope conductance of 215

33 pS (N 5). The average permeability ratio of K- to Na-,

calculated from inside-out patches with Na* patch solution on the

extracellular side of the patch and KC1 saline on the intracellular

side, was 3.9 0.9 (N = 9). Increasing the Ca*- concentration

bathing the internal face of this channel from 10-7 M to 10-6 M or

10-5 M increased the frequency of opening (Fig. 3-10). At a

constant internal Ca"* concentration, depolarization also increased

the Po by increasing the mean open time, with a weak tendency to

decrease the mean closed time as well (Fig. 3-9B). Neither 20 mM

TEA, 5 mM 4-aminopyridine, nor 10 mM Cs- in KC1 saline blocked this

channel when applied to its intracellular side.

Fig. 3-9. Plots of the I-V relationship and kinetic properties of a
Ca---activated potassium channel. A. The current-voltage
relationship of a single channel in an inside-out patch in
symmetrical 480 mM KC1. The slope conductance was 197 pS. B. The
mean open (open circles) and closed (closed circles) times versus
voltage. Na+ saline containing 10-5 M Ca*+ has been substituted as
the intracellular solution. The curves were fit to the data by
nonlinear regression.



r -5-


-75 -50 -25 0 25 50

5.0 5.0


z 2.5 -2.5 0
z o
L z

0.0 0.0
-50 -30 -10 10


Fig. 3-10. Representative records showing the Ca-* dependence of
the activation of a 215 pS K- channel in an inside-out patch at -10
mV. Openings became longer and more frequent as the Ca"
concentration was increased from 10-7 M (top trace) to 10-6 M
(bottom trace). Na- patch solution was in the pipette and KC1
saline was the intracellular solution. Arrows indicate the current
level of the closed state.

1 4pA

-- iw~L, 311lp.rCg~.i~,4-r~r+r~fnl(h3r

e- ^ ^ 1^} I

A voltage-dependent K' channel

Depolarizing voltage steps applied to cell-attached patches

often revealed a channel carrying outward current. Hyperpolarizing

prepulses of -30 mV or -50 mV increased the frequency of channel

openings but were not a necessary condition for activation of the

channel. With Na* salines in the patch pipette, the slope

conductance for outward currents was 9.7 2.6 pS (N = 3). The

slope conductance of the inward current through the cell-attached

channel was 49.4 14.8 pS with KC1 saline in the patch pipette (N

= 5). The inward rectification predicted by these slope

conductances was small but apparent in the I-V relationships (Fig.

3-11A, B). Reversal potentials (in mV from the resting membrane

potential) extrapolated from the most linear portions of I-V

relationships were 45 14 mV (N = 4) with 480 mM KC1 applied

extracellularly and -105 14 mV (N = 3) with 480 mM NaCl applied

extracellularly. These values suggest a selectivity for K+ over

Na*. The activation of this channel by depolarizing voltage pulses

was delayed. The first latencies of one channel activated by

depolarizing pulses preceded by hyperpolarizing prepulses was 259

147 ms at -25 mV from rest, 55 89 ms at 0 mV from rest, and 27

15 ms at 25 mV from rest. Channel activity sometimes occurred

throughout the duration of depolarizations lasting as long as 5 sec.

However, steady-state activity was never observed, leading to the

conclusion that strong inactivation occurred, but with a slow onset.

In inside-out patches held at -70 mV, channel activity was observed

Fig. 3-11. A voltage dependent, K selective channel with
properties of a delayed rectifier. A. Representative recordings of
a channel in a cell-attached patch during activation by 5 s voltage
steps from a holding potential of -30 mV from rest. The pipette
contained Na- patch solution. Arrows indicate the current levels of
the closed state. fc 1.5 kHz. B. The I-V relationships of two
other channels in cell-attached patches with 480 mM KC1 (open
circles) and 480 mM NaCl (closed circles) in the pipette. Note the
inward rectification displayed by this channel. The curves were fit
to these data by nonlinear regression.


72 5 N '' t /' 1


25 J1? 74' 9^/J4^

0- r i H |2pA








-60 -20

20 60 100 140


during voltage steps to -60 mV or above. Rare openings at -70 mV

did occur at the offset of depolarizing voltage steps.

A steady state Cl- channel

Also observed in cell-attached patches was a 35 13 pS channel

whose reversal potential was near the resting membrane potential

(-8.2 mV 9.5 mV from rest, N = 9). This channel was relatively

common, being found in 26 of 102 patches, and was selective for Cl-

over cations. The current-voltage relationships (Fig. 3-12A, B) of

inside-out patches with 480 mM KC1 on the extracellular face of the

patch and 480 mM NaCl on the intracellular face, gave a mean

reversal potential of -2.0 1.8 mV (N = 5). In these inside-out

patches the slope conductance was 70 4 pS. In three channels in

inside-out patches, intracellular perfusion with the Cl- channel

blocker potassium isothiocyanate (20 mM) reduced the slope

conductance of the chloride channel by 45Z and 57Z, while the third

channel was unaffected.

The Cl- channel was active at all holding potentials tested and

its Po was relatively independent of the holding potential (Fig. 3-

13). The distributions of the open dwell times were usually best

fit by single exponentials, giving mean open times ranging from 1.1

to 6.9 ms. The distributions of the closed dwell times were better

fit by two exponentials, with mean closed times ranging from 1.6 to

3.9 ms. Neither the mean open or closed times showed any

consistent dependence on membrane potential.

Fig. 3-12. A steady state, Cl- selective channel. A.
Representative recordings of a Cl- channel in an inside-out patch
with KC1 saline in the patch pipette and an intracellular solution
of NaCl saline. Arrows indicate the current level of the closed
state. f. = 1.5 kHz. B. The I-V relationship of a another Cl-
channel in an inside-out patch under the same conditions. The
slope of the linear regression fit to these data was 71 pS.

50V 4.0
LJ |

1 2.0

Z 0.0


-60", r -4.0
-60 -30 0 30 60


Fig. 3-13. Plot of the Po of a Cl- channel versus voltage. The Po
of this channel displayed little dependence upon voltage in either
the cell-attached (closed circles) or the inside-out (open circles)
configuration. These data were obtained from the same patch before
and after excision. Lines were fit to the data by linear






0 _- ---- w-



-40 -10




i n












I i i i I I I I



The nonselective cation channel

The nonselective cation channel shares several properties with

a channel found in Aplysia neurons (Chesnoy-Marchais, 1985): 1) it

is resolved only in inside-out patches; 2) high internal

concentrations of Na, but not Cs* or K*, appear to prevent

inactivation; 3) its kinetics are sensitive to the type of divalent

cation present; 4) it is weakly voltage dependent; 5) it appears to

have a multi-ion pore, and if it is permeable to divalent cations as

my data suggest; 6) Ba** gives a higher conductance than Ca**; and

7) its current-voltage relationship shows inward rectification in

high external concentrations of divalent cations, especially at

potentials above EN..

Mg** block offers a partial glimpse of the structure of the

cation channel. The channel appears to have at least two binding

sites for cations in its pore, and another divalent cation binding

site outside the electric field of the membrane capable of altering

the gating properties of the channel. This latter site may be very

near the permeation pathway, however, since both internal and

external permeable divalent cations decreased the Po.

Alternatively, a second, external binding site could exist. In

addition to agreeing with the Woodhull model of channel block, the

interpretation that fast block was due to Mg-" entering the pore is

consistent with evidence that Na- and Mg-* compete for binding

sites, namely that the dissociation constant for Mg** depended on

the direction of flow of Na* through the pore. When the holding

potential was above EN. at 0 mV, where Na should force internal

Mg-- into the channel, the apparent dissociation constant was in the

millimolar range. When the holding potential was below EN., where

Na* should oppose the entry of internal Mg*" into the pore, the

apparent dissociation constant was decreased 100-fold. This type of

interaction between permeant ions is consistent with a multi-ion

pore, but to prove this conclusion, future studies would need to

show that Na* causes an increase in the dissociation rate constant

of Mg- (Moczydlowski, 1986). Further evidence for a multi-ion pore

could also be obtained by investigating the mole fraction dependence

characteristic of multi-ion pores (Hille and Schwarz, 1978). My

data would predict a conductance minimum at about 10 mM divalent

cation and 480 mM Na+, a ratio of about 0.2.

If the channels permeable to divalent cations in inside-out

patches were, in fact, nonselectiv* cation channels, then the

divalent cations replaced Na- as the major permeant ion. Ca-- or

Ba- can supersede monovalent cations as major permeant ions in

channels when they bind more strongly to sites within the channel

pore, effectively excluding the monovalent cation (Dani and

Eisenman, 1987). A necessary corollary of this assertion is that

Ca'* and Ba-* should have a lower conductances than Na-. While my

data do not directly address this point, Ba-", which appeared to

bind less strongly within the pore than Ca-- by virtue of its weaker

blocking ability, gave a higher conductance than Ca-* under

identical conditions. Furthermore, the properties of the fast block

by Mg-* strongly suggest that divalent cations bound more strongly

to sites within the pore than monovalent cations.

Internal divalent cations caused reductions in the P. by

interacting with other sites apparently outside the electric field

of the membrane. The effect appeared to be mediated by a site (or

sites) on the channel protein and not by charge screening of the

diffuse negative charges on the membrane (which alters the surface

potential and effectively changes the profile of potential across

the membrane). Charge screening by internal cations should cause a

negative shift in the activation curve of voltage dependent channels

(McLaughlin et al., 1971), whereas I observed a positive shift.

K* channels

The large conductance and Ca-1 dependency of the Ca-activated

K+ channel is similar to channels from several types of vertebrate

cells (Wong and Adler, 1986; Yellen, 1984; Pallotta et al., 1981).

A similar 130 pS Ca---activated potassium channel has been reported

in mouse olfactory receptor neurons (Maue and Dionne, 1987). The

mouse and lobster channels are similar in their Ca-- concentration

dependency, resistance to TEA, and mean open durations, but differ

in their sensitivity to Cs- block.

The voltage-activated K* channel in lobster olfactory receptor

cells is most similar to the delayed rectifier class of K- channels

because of its delayed activation, slow inactivation and small

unitary conductance. Although its increased activation following

prehyperpolarization and its inward rectification are not properties

of classical delayed rectifier currents, they are not inconsistent

with the heterogeneity among channels now classified as delayed

rectifier channels (Rudy, 1988). A similar channel, but with a much

larger slope conductance of 40 pS, was found in mouse olfactory

receptor cells (Maue and Dionne, 1987). The channels from mouse and

lobster olfactory receptor cells both show slow inactivation and

similar first latencies.

Chloride channel

The steady state Cl- channel appears to provide the receptor

cell with a constant leak pathway for Cl-, suggesting that the

lobster olfactory receptor cell has a passively distributed

concentration gradient of Cl-. That the channel's reversal

potential in cell-attached patches was close to the resting membrane

potential agrees with this conclusion. Block of the channel by

isothiocyanate, which is a permeant anion in some Cl- channels

(Franciolini and Nonner, 1987), agrees with the isothiocyanate block

of the voltage-gated Cl- channel from Torpedo electric organs (Tank

et al., 1982). A voltage-dependent Cl- channel of larger

conductance has been described in mouse olfactory receptor neurons

(Maue and Dionne, 1987).

Significance for olfaction

The two K* channels in lobster olfactory receptor neurons

appear to have counterparts in vertebrate olfactory receptor cells

(Maue and Dionne, 1987; Firestein and Werblin, 1987; Trotier, 1986);

all probably function in membrane repolarization. They probably

also underlie the outward rectification observed in lobster

olfactory receptor cells (Schmiedel-Jakob et al., 1989), which may

help to limit the magnitude of odor-evoked depolarizations to the

voltage range over which the cell is capable of producing spikes.

Unlike mouse and salamander olfactory receptor cells, however,

lobster receptor cells do not appear to have inward rectifier K-

channels, believed to contribute to the stability of the membrane

potential (Maue and Dionne, 1987; Trotier, 1986). This function

could be subserved in lobster olfactory receptor cells by the steady

state Cl- channel. This Cl- channel may also be partly responsible

for the lower input resistance of lobster olfactory receptor cells

(Schmiedel-Jakob, 1989) compared to that of salamanders (Firestein

and Werblin, 1987; Trotier, 1986).

The function of the cation channel, however, remains a mystery.

Conceivably, it could enhance the spread of odor-evoked

depolarizations by summing its inward current with the odor-evoked

inward currents, effectively increasing the space constant of the

cell (Yoshii et al., 1988). Positive feedback would be avoided if

the nonselective cation current was less than the outward leak

current. The lack of any convincing evidence that it is active in

intact cells and its suppression by Mg"*, however, suggest that the

channel is unlikely to be active under normal physiological

conditions. Similar channels in molluscs appear to be activated

only by physiologically stressful conditions. Strong et al. (1987)

observed this type of channel in cell-attached patches from Aplysia

neurons, but only following injury to the processes of these cells.

Yazejian and Byerly (1989) recently recorded a similar channel in

Helix neurons and found that it could be activated by sustained


intracellular perfusion with high Ca-* solutions or low Ca-*

solutions lacking ATP, or by a low Ca"*, high K* bath solution.

Since my data come from cells bathed in low divalent saline, which

is probably stressful to the cells, it is possible that this species

of cation channel is also involved in responses to trauma in the

lobster. Trauma-induced resorption of processes might occur

regularly in neurons whose dendrites are exposed to damage as are

those of olfactory receptor cells.