Rhythmic Activities of Isolated and Clustered Pacemaker Neurons and Photoreceptors of Aplysia Retina in Culture

Jon Jacklet, ',* Steven Barnes, Andrew Bulloch, Kenneth Lukowiak, and Naweed Syed2

'Department of Biological Sciences, State University of New York at Albany, Albany, New York 12222 and 'Neuroscience Research Group, University of Calgary, HSC, 3330 Hospital Drive N.W., Calgary, Alberta T2N 4N1. Canada

SUMMARY

Each eye of Aplysia contains a circadian clock that pro- duces a robust rhythm of optic nerve impulse activity. To isolate the pacemaker neurons and photoreceptors of the eye and determine their participation in the circadian clock and its generation of rhythmic autoactivity, the ret- ina was dissociated and its cells were placed in primary cell culture. The isolated neurons and photoreceptors survived and vigorously extended neurites tipped with growth cones. Many of the photoreceptors previously de- scribed from histological sections of the intact retina were identified in culture, including the large R-type pho- toreceptor, which gave robust photoresponses, and the smaller tufted, whorled, and flared photoreceptors. The pacemaker neurons responsible for the rhythmic impulse activity generated by the eye were identified by their dis- tinctive monopolar morphology and recordings were

made of their activity. Isolated pacemaker neurons pro- duced spontaneous action potentials in darkness, and pacemaker neurons attached to fragments of retina or in an isolated cluster interacted to produce robust sponta- neous activity. This study establishes that isolated reti- nal pacemaker neurons retain their innate autoactivity and ability to produce action potentials in culture and that clusters of coupled pacemaker neurons are capable of generating robust autoactivity comparable to pacemaker neuron rhythmic activity recorded in the intact retina, which was previously shown to correspond to 1:l with the optic nerve compound action potential activity. o 1996

Keywords: circadian rhythm, cell culture, pacemaker neurons, autoactivity, action potentials.

John Wiley &Sons, Inc.

INTRODUCTION

The paired cephalic eyes of the marine gastropod Aplysia are small and inconspicuous, but each eye contains a circadian clock that produces a robust and precise circadian rhythm of impulse activity. All the essential components required for the clock are present in an isolated eye including the basic circadian pacemaker, photoreceptors to entrain

Received January 24, 1996; accepted April 1 I . 1996 Journal of Neurobiology, Vol. 3 I , No. 1, pp. 16-28 ( 1996) 0 1996 John Wiley & Sons, Inc. CCC 0022-3034/96/0 10016-13

JWJ74@cscpop.albany.edu).

16

*To whom correspondence should be addressed (e-mail:

the pacemaker, and an output pathway that com- municates the clock information to the central ner- vous system (CNS). The output axons enter the optic nerve (ON) and project into the CNS along well-defined pathways that carry the impulse activ- ity (Olson and Jacklet, 1985; Strack and Jacklet, 1993). Rhythmic behaviors such as the circadian rhythm of activity/rest (Strumwasser et al., 1979; Lickey et al., 1977) are controlled by the ocular cir- cadian clock. It also controls some features of pho- toresponsiveness of the eye itself ( Jacklet, 199 1 ) .

The cellular organization of the circadian clock in the Aplysia eye has been determined in part (review, Jacklet, 1989), but important questions remain unanswered about the function and capa- bilities of individual pacemaker neurons and pho-

Aplysia Retinal Pacemaker Neurons 17

toreceptors within the eye. Is an isolated pace- maker neuron capable of producing action poten- tials? Do either specialized photoreceptors or pacemaker neurons respond to light? Do coupled neurons generate more robust and precise autoac- tivity than an isolated neuron does? Once these initial questions are answered and the necessary conditions for long-term recording from cultured neurons are established, the larger questions con- cerning the ability of isolated or clusters of neurons to generate a circadian rhythm and the cellular mechanisms of the rhythm can be addressed.

A circadian rhythm in the frequency of com- pound action potentials (CAPS), can be recorded from the ON of an isolated eye (Jacklet, 1969a), and the rhythm persists for a week or more if the isolated eye is maintained in darkness and organ culture. CAP activity is produced by the synchro- nous autoactivity of a population of retinal neu- rons, now called pacemaker neurons (PNs) be- cause they provide the pacemaker activity that drives the CAP frequency (Jacklet, 1989). The PNs had previously been called secondary neurons, because they were thought to receive serial in- put from photoreceptors ( Luborsky-Moore and Jacklet, 1977; Jacklet et al., 1982), and neurosecre- tory cells (Strumwasser et al., 1979; Herman and Strurnwasser, 1984), because they contain dense- core synaptic vesicles ( Luborsky-Moore and Jacklet, 1977; Strumwasser et al., 1979). The PNs were identified by intracellular recording and dye injection (Jacklet et al., 1982). Their cell bodies are spherical and 15-23 pm in diameter. They are monopolar and contain secretory inclusions re- sembling lipochondria ( Luborsky-Moore and Jacklet, 1977). PNs usually appear in small clus- ters and are confined to the outer layer of the retina near the origins of the ON (Herman and Strum- wasser, 1984). Monopolar neurons, identical to the PNs, were isolated from the retina and placed in culture by Strumwasser et al. ( 1979), but their ability to produce action potentials was not re- ported.

The retina of a related mollusk, Bulla, also con- tains a circadian clock (Block and Wallace, 1982) and the basal retinal neurons (BRNs), similar to the PNs of Aplysia, generate the CAP activity rhythm ( McMahon et al., 1984). BRNs were iso- lated in cell culture and shown to express a circa- dian rhythm of membrane conductance change (Michel et al., 1993). However, isolated BRNs were not yet shown to exhibit spontaneous action potentials in darkness, so the question of whether they are capable of producing a circadian rhythm

of action potential rate is unanswered (Michel et al., 1993). Here we show that isolated and clus- tered Aplysia PNs do retain the ability to produce spontaneous action potentials in darkness in cul- ture.

A variety of photoreceptors (PRs) should ap- pear in Aplysia retinal cell cultures because the ret- ina contains several classes of PRs in addition to PNs. The largest (30-50 pm) and most conspicu- ous PR is the R type (Jacklet, 1969b). It is depolar- ized by light in proportion to light intensity (Jacklet and Rolerson, 1982) and has a conspicu- ous microvillus photoreceptive process. The R- type PRs and others have axons in the ON in par- allel with PN axons (Jacklet and Rolerson, 1982). Four other smaller PR types that have either mi- crovillar or ciliated photoreceptive processes were revealed by detailed microscopical analysis of the eye ( Herman and Strumwasser, 1984).

In this study we dissociated the retinal cells, maintained them in primary culture, observed their outgrowth, and made whole-cell patch re- cording of their membrane potentials, action po- tential activity, and ionic currents. We determined that isolated PRs, PNs, and PN clusters are capable of generating activities similar to those recorded from the retina of the intact eye.

MATERIALS AND METHODS

Eyes were dissected from Aplysia caljfornica, weighing 75- 150 g, obtained from either Marinus (Long Beach, CA) or Aplysia Resource Center (Miami, FL), and maintained in Calgary at 16°C on a light and dark sched- ule ( I2 h each) in seawater aquaria. Eyes from more than 100 animals were used.

Cell Culture

Cells were dissociated from the retina and plated in cul- ture dishes using standard techniques (Schacher and Proshansky, 1983; Kleinfeld et al., 1990; Syed et al., 1990; Ridgway et al., 199 1 ). Culture medium was essen- tially as described by Kleinfeld et al. (1990) and con- sisted of an isotonic saline (see below, Whole Cell Recordings) and L- 15 powder mix without glutamine and organic salts (No. 82-5 154EA; Gibco, Grand Island, NY). Eyes were treated with 1% protease (type IX, Sigma) in L- 15 medium for several hours at 33 or 36"C, rinsed, and then cleaned of connective tissue. The lens was removed and the retina was cut near the equator of

CORNEA

HORIZONTAL MERIDIAN

CORNEA

HORIZONTAL CUT MERIDIAN

EQUATOR

Figure 1 Diagram of an eye of Aplysia. The salient fea- tures are shown including the cornea, the equator, the horizontal meridian, the central pole (opposite the cornea), retinal fiber tracts that converge to form the or- igins of the ON, and the basal retinal area (crosshatch) where PNs are found. The eye was cut (between the arrows) and the corneal portion discarded to reduce the eye to the basal retina prior to dissociation of the cells. The diagram is adapted from Herman and Strumwasser ( 1984).

the eye (Fig. 1 ). To increase the proportion of PNs in culture the corneal portion of the retina was discarded and the basal retina containing the O N head and the area near the central pole, which contains the majority of the PNs (Herman and Strumwasser, 1984)- was retained. The basal retina is thick and contains many PRs in the inner retinal layer interdigitated with pigmented cells ad- jacent to the lens (Jacklet and Rolerson, 1982). Axons of PRs and PNs extend into the outer layer of the retina where they form the fiber tracts that combine and be- come the ON (Fig. 1 ). The basal retina was minced and then triturated gently in filtered (0.45-pm) Aplysia hem- olymph using a 10-pL pipette tip. Dispersed cell were plated on poly-L-lysine (Sigma, >500,000 M W ) coated Corning 25,000 culture dishes containing L-15 culture medium. Culture medium consisted of 50% modified L-15 medium made up using an isotonic saline appro- priate for Apl~vsia (Kleinfeld et al., 1990) and 50% fil- tered (0.45-pm) Aplysiu hemolymph. Plated cells were maintained in dim room light at room temperature until recordings were made, usually on the first o r second day after isolation. Culture medium was changed sometimes after initial plating and replaced with L-15 medium alone, without hemolymph. Photographs of plated cells were taken using a Zeiss inverted microscope equipped with phase contrast optics or a Nikon Diaphot equipped with Hoffman contrast optics.

Whole Cell Recordings Whole-cell tight-seal recordings were made with 1-5 megaohm ( M Q ) electrodes filled with either 5 I I m M

KCI or 51 1 m M K gluconate in addition to 2 m M MgCI2,20 mMK-HEPES, 1.5 mMEGTA,and I.OmM CaClz at pH 7.6. We used ruptured patch or nystatin per- meabilized patch techniques (final nystatin concentra- tion I50 pg/mL). The cells had capacitances of 9.5 f 3.3 pF ( n = 16, M k S.D.) and input resistances of 6.5 f 6.6 GO, ( n = 16, M f S.D.). Cells were superfused with sa- line containing ( in m M ) : 460 NaCI, 10 KCI, 10 CaCI2, 50 MgCI2, and 20 HEPES, adjusted to p H 7.8. Bath ref- erence electrode was an agar bridge with a chlorided sil- ver wire immersed in the bridge holder. Neurons were observed using a Nikon Diaphot equipped with Hoffman contrast optics and recordings were made using an Axo- patch 200 in either current or voltage clamp mode. Volt- age clamp data (sampled at 100 Hz) and current clamp data (sampled at 200 Hz) were written to disk using BA- SIC-FASTLAB ( Sunnyvale, CA) routines. Illumination was provided by a 100-W halogen microscope lamp, ad- justed in intensity, or white fluorescent light. Recordings were made at room temperature ( 19-2 1 "C) and the mi- croscope light was heat filtered. Recordings lasting < l h were made from > 100 cultured cells.

RESULTS

Morphology of Isolated PNs and PN Clusters In Vitro

The degree of dissociation of the retinal cells was determined by the amount of trituration used dur- ing the dissociation, and thus we could obtain a mixture of isolated cells, clusters of cells, and frag- ments of retina according to our needs. All these types of cells and tissues became attached to the culture dish and, except for pigmented support cells, began to extend neurites tipped with growth cones.

Isolated PNs had distinctive morphology in cul- ture [Fig. 2(A)] , as they do in the intact retina. The cell body was spherical, 15-23 in diameter, and contained a few dark inclusions. A single slen- der neurite extended from the cell body and no other processes appeared on the cell body, in con- trast to other cells of a similar size that had an api- cal process or photoreceptive tuft or an oblong cell body with a thick initial neurite [see Fig. 9( C,D)] . After less than 12 h in culture, the single slender neurite emerged from the cell body and continued to grow for several days. The tip of the growing neurite had a large motile growth cone with distinc- tive lamellipodia and filopodia [Fig. 2 (B)] .

Some PNs were not completely dissociated but appeared in clusters of two to six, similar to the clustering of PNs seen in the intact retina. Figure 2( C ) shows a cluster of three PNs that appear to be

Aplysia Rrrinal Pucrrnaker Neurons 19

Figure 2 Dissociated PN and PN cluster. ( A ) An isolated PN with an outgrowing neurite and active growth cone. ( B ) An enlarged view of the growth cone in (A) . (C) A cluster of three PNs joined at their cell bodies. The tiny cell on the lower edge of the cell body cluster is an unknown type. Phase contrast optics were used. Scale bar = 20 fim (A , C).

attached at their somata; each PN has a short neu- rite tipped with a growth cone.

Often a PN was attached to a retinal fragment by its intact axon only (Fig. 3 ) . These PNs had the same morphology as isolated PNs: spherical cell body lacking processes other than a slender axon. The approximate position of the PN and the frag-

ment in the intact retina could be determined by the type and position of cells in the fragment. That is, PNs are found in the outer retinal layer and their axons project into the neuropil and fiber tracts of the eye, and pigmented cells are found only in the inner retinal layer associated with the distal photo- receptive segments ofthe PRs. Thus in Figure 3 the

20 Jucklet et ul

Figure 3 Partially dissociated retinal fragments and attached PNs. ( A ) A partly dissociated retinal fragment with a PN attached by its axon to the fragment, which is composed of PRs, pigment cells, and other PNs. ( B ) Another larger fragment with attached PN and membrane patch recording electrode on the PN. Scale bar = 40 pm.

presence of attached pigmented cells identified PRs derived from the inner retinal layer and the PN likely came from the outer layer. The PNs that were attached to retinal fragments provided the oppor- tunity for us to record their electrical activity and determine if PNs retained their ability to interact with neighboring PNs and PRs (see below).

Electrical Activity of Isolated PNs

Isolated PNs, e.g., the one with a neurite and growth cone in Figure 2 (A) , had resting potentials in the range -35 to -40 mV (Fig. 4) . Typically, the membrane potential exhibited more frequent slow depolarizing fluctuations [Figs. 4 3 (A)] than the potential of an attached PN [see Figs.

T o Y 2o i ’ -40 -20 t

4 s

5 ( B ) ,6,8 1. An isolated PN was capable of produc- ing spontaneous overshooting action potentials (Figs. 4,5), if it had extended a neurite. PNs lack- ing neurites were sampled ( n = 13 ), and although the resting potentials were about -40 mV, they never produced action potentials, even in response to applied depolarizing current or depolarization following hyperpolarization. In agreement with the absence of action potentials, PNs lacking neurites also lacked early inward currents when voltage clamped. Action potentials in isolated PNs with neurites arose from slow spontaneous membrane depolarizations [Figs. 4,5 (A)] and lacked the smoother pacemaker potentials and subthreshold pulsatile depolarizations characteristic of attached PNs [see Figs. 5(B),6,8]. Longer term recording from PNs [Fig. 5 (A)] show that an isolated PN has somewhat erratic firing rates but average about 0.5 Hz. Twelve of 16 isolated PNs expressed spontane- ousactivity in darkness at 0.05-0.5 Hz, and all pro- duced action potential when depolarized with cur- rent.

Electrical Activity of PN Clusters

Figure 4 Spontaneous action potentials recorded in darkness from an isolated pacemaker neuron like the one shown in Figure 2( A). Membrane potential is about -35 mV; action potentials are 55-60 mV in amplitude.

Clusters containing two to six closely associated PN somata were commonly found after Plating. This appears to be the result of the innate clustering of PNs in the retina. The junctions between somata

Aplysia Retinal Pacemaker Neurons 21

A

0 " e

Figure 5 Spontaneous action potential activity recorded in darkness. (A) Activity from an isolated PN like the one shown in Figure 2(A). Membrane potential is about -35 mV; action potentials are 50-60 mV in amplitude and occur at about 0.5 Hz. Membrane potential irregu- larity is caused by frequent slow depolarization. ( B ) Activity from a cluster of two PNs similar to the ones shown in Figure 2( B ) . Membrane potential is about -40 mV; action potentials are about 50-60 mV in amplitude and occur at about 1 Hz.

of PNs are apparently strong enough to prevent the complete dissociation of PNs in a cluster. PNs at- tached to a cluster had resting potentials near -40 mV. They expressed robust autoactivity in dark- ness consisting of pacemaker potentials giving rise to overshooting action potentials 60-70 mV in am- plitude [Fig. 5(B)]. The action potential fre- quency of a PN in each of eight clusters was exam- ined. Frequencies were relatively constant ( range 0.05-1 .O Hz) and less erratic than the frequency of isolated PNs. Pulsatile subthreshold depolariza- tions often preceded the action potentials recorded from PNs in a cluster (Fig. 6) . Similar pulsatile de- polarizations occurred in recordings from PNs in

> o / . - . . . .

I I

4 s

Figure 6 Spontaneous action potentials recorded in darkness from a PN attached to a cluster of six PNs, sim- ilar to the cluster shown in Figure 2 (B). The membrane potential was depolarized from rest (-40 mV) by a pace- maker potential and action potentials were evoked. Ac- tion potentials are often preceded by pulsatile depolar- izations, similar to those seen in PNs attached to retinal fragments.

the intact retina and are believed to be caused by electrotonic interactions (Jacklet et al., 1982). Such interactions are also seen in PNs attached to retinal fragments (see below).

Attempts to voltage clamp the membrane of a PN attached to a cluster revealed fast inward cur- rent pulses associated with undamped action po- tentials that presumably occurred in remote but coupled neurons. In favorable circumstances ( n = 9) voltage control was obtained and the ionic currents of the PN were recorded (Fig. 7). Using a voltage clamp paradigm in which the voltage was held at -30 mV and a series of steps between - 120 and +40 mV were applied, three types of ionic cur- rents were observed: a large transient inward cur- rent, a late outward current evoked by depolariza- tion above -30 mV, and a transient outward cur- rent evoked by the depolarization to -30 mV after a preceding hyperpolarization. This current is largest when evoked by the step from - 120 to -30 mV and becomes progressively smaller with smaller hyperpolarizations [Fig. 7 (A)]. The cur- rent-voltage relationship [ Fig. 7 ( B)] reveals that the transient inward current and late outward cur- rent (likely responsible for depolarization and re- polarization of the action potential) are activated positive to -20 mV. The transient outward current evoked by returning the membrane voltage to -30 mV, after the membrane voltage hyperpolarizing pulses [Fig. 7(A)], is characteristic of A current, which is associated with pacemaker potentials in mollusks. This group of PN currents is distinctive

22 Jacklet et al.

-1 00 1 11

0

B

100 200. 300 time (ms)

400

500 T f

500

Figure 7 Currents and current-voltage relationship taken from a voltage clamped PN that exhibited sponta- neous action potentials and was attached to a cluster of five other PNs. ( A ) Currents recorded in response to voltage steps from holding voltage, -30 mV. Clamp steps were in the range - 120 to +30 mV in 10-mV incre- ments. Three current types are apparent: a transient in- ward, late outward, and transient outward A current. A current is evoked by the step back to -30 mV at the end o f a hyperpolarizing pulse, in the range ~ 120 to -70 mV. ( B ) Current-voltage relationship of the PN membrane. A current (squares) is largest upon return to -30 mV from ~ 120 mV. Peak transient inward (circles) and late outward (triangles) currents are activated positive to -20 mV. Note the long decay time of the transient out- ward A current; and the current in response to +40 mV, shown in ( B) , is not represented in ( A ) because the cur- rent was off scale and required a gain change.

and quite different from currents recorded from PRs, which lack A current and substantial voltage- activated inward current (see Fig. 1 1 ).

Electrical Activity of Fragments Containing PNs and PRs

Often a retinal fragment containing PNs, PRs, and pigmented support cells had a PN attached to it in

a favorable position to record its activity (Fig. 3 ) . We took advantage of this and recorded from a PN if it was attached only by its axon. Thus, we could determine the kind and level of activity that was produced by a PN receiving input from other PNs and PRs in the attached retinal fragments.

The resting potentials of a PN attached to a reti- nal fragment containing PNs and PRs was about -40 mV (the same as an isolated PN). Most PNs exhibited overshooting action potentials 60-70 mV in amplitude (Fig. 8). They occurred sponta- neously in darkness [Fig. 8 ( A,B,C)] and their fre- quency was increased by illumination [Fig, 8 (D,E)] . Pacemaker potentials preceded the ac- tion potentials and often small pulsatile depolar- izations [Fig. 8( B,C,D)] contributed to the depo- larization, leading to the overshooting action potentials. The pulsatile depolarizations were im- perfectly synchronized and might occur on either the leading or trailing edge of the overshooting ac- tion potentials. They appeared to originate from neighboring neurons in the fragment or cluster be- cause they were only observed in attached PNs. Sometimes they were not large enough to evoke an action potential and thus appear as subthreshold pulses [Fig. 8(A,D)]. When a fragment was ex- posed to light, the PN was depolarized and the fre- quency of overshooting action potentials increased while the frequency of pulses decreased [Fig. 8 (E)] , suggesting that a pulsatile depolarization combined with a light-induced depolarization could reach the threshold and evokes an overshoot- ing action potential. Clearly, simultaneous record- ings from pairs of neurons are needed to be certain of the hypothesized interactions. Action potential frequency remained relatively constant through the duration of the illumination, even over 10- 100 s [Fig. 8( E)] . This is characteristic also of PN re- sponses driven by light in the intact retina (Jacklet, 1969b; Jacklet et al., 1982).

Isolated PRs

Many retinal cells in dissociated cell culture were classified as PRs because they had apparent photo- receptive processes on their cell bodies similar to those described in histological sections of the intact retina. The largest and most conspicuous was the microvillus or R type [Fig. 9(A)]. It usually had an oval shape and a microvillus process at the apex. Typically, several short thin neurites extended from the cell body. The R type was often associated with other PRs and pigmented support cells [Fig. 9( B)] in an incompletely dissociated fragment.

Aplysia Retinal Pacemaker Neurons 23

40 r A B

IOt&> -40

C I

--- I s 1 s I s

D l

1 I

3s

light (50 s)

Figure 8 Activity recorded from a PN attached to a retinal fragment like the one in Figure 3 ( B ) . ( A-C) Spontaneous action potentials recorded in darkness and (D) recorded in light. A variety of small and partially synchronized, subthreshold pulsatile depolarizations are associ- ated with each large overshooting action potential. ( E ) Similar activity in darkness that in- creases in frequency in response to a light stimulus. Light depolarized the neuron and evoked a train of action potentials that are full sized except for pulses two and three. The first three pulses in response to light in ( E ) are shown in detail in (D) . White light, 70 pW/cm2. Data sampled at 2 kHz, and each action potential was 2 ms at half amplitude, giving adequate sampling at 3 to 4 points for each action potential.

Pigmented support cells were darkly pigmented and spherical and did not sprout processes in cul- ture. Because they always appear to be spherical in culture, their columnar shape in the intact retina is likely dictated by association with other retinal cells in the intact retina. Many of the other smaller PRs observed in culture retained a shape more characteristic of their shape in the retina. The PR shown in Figure 9 ( C ) has a spindle shape and an obvious photoreceptive distal segment and most closely resembles a tufted or whorled PR, described from the intact retina by Herman and Strumwasser ( 1984). Frequently a cell type was observed that appeared to be a small PR be- cause it had a stout axon hillock and an apparent photoreceptive process at the apex, which, never- theless, had a short process actively extending from its apex [Fig. 9 ( D ) ] . This cell type may be the precursor to bipolar cells seen in culture that had a neurite growing from each end. Such cells were often seen in cultures that were more than several days old. We did not attempt to classify the smaller PRs and their membrane activities in further detail.

We recorded the photoresponses of the PRs with the large microvillus R-type morphology

[see Fig. 9 (A ,B) ] to confirm their identity. PR photoresponses were graded in proportion to light. A short ( 1-s) bright light pulse caused a prolonged depolarization that outlasted the light pulse and approached the 0-mV membrane po- tential level before declining to the resting mem- brane level [Fig. 1 0( A) ] . A bright and long (26- s ) light pulse caused the PR depolarization to re- verse and become a long-lasting hyperpolariza- tion [Fig. 10( B) ] . These responses are similar to photoresponses of R type PRs, identified by dye injection, obtained from the intact retina (Jacklet and Rolerson, 1982) . Other PRs in the intact retina gave complex light responses con- sisting of hyperpolarization or hyperpolarization combined with subsequent depolarization and spike activity (Jacklet and Rolerson, 1982 ).

Three obvious R-type PRs, with pigmented photoreceptive processes, that responded to light with a graded depolarization were voltage clamped and currents evoked in response to a se- ries of voltage steps using the same paradigm used to test PNs (see Fig. 7). Their currents, shown in Figure 11 were rather different from those of PNs, because they lacked substantial early inward current and A current.

24 Jacklei et al.

Figure 9 Types of PRs and associated pigment cells in culture. ( A ) An isolated microvillus or R-type PR with short microvilluslike projections (arrow) from the apex (toward the left) and a small thin neurite and associated growth cone attached to its base. ( B ) Two R-type PRs (apex upward) in association with five pigment cells. Small thin basal neurites tipped with growth cones project from the PR base. (C) A small PR (possibly a tufted or whorled PR). It has a distinctive spindle-shaped cell body and a short apical photoreceptive process (arrow). ( D ) Another small PR (possibly a flared PR) with an apparent process growth from the apex (arrow). Scale bar = 20 pm ( A-C).

DISCUSSION

All cells obtained by dissociating the retina, except the pigmented support cells, extended processes within 12 h of plating. Neurite outgrowth was guided by active growth cones like those described for molluskan neurons by Schacher and Proshan- sky ( 1983). Most of our cultures were used within 2 days of plating for recording, but some were kept

for 4-5 days. In those cultures, outgrowth contin- ued for several days from most of the PNs and PRs and quite elaborate fiber interactions, including fasciculation (not shown), took place.

It was necessary to have Aplysia hemolymph in the culture medium to obtain prompt and vigorous neurite growth, as noted by others for central neu- rons of Aplysia (Schacher and Proshansky, 1983). Hemolymph may contain both growth promoting

Aplysia Retinal Pacemaker Nwrons 25

Y

> I IT. -25 1 n r )

U light (1 s)

or . . .

-50 L I I light (26 s)

Figure 10 Current clamp recordings from an isolated R-type PR kept in darkness for 10 min prior to illumina- tion. ( A ) Response to a 1 -s light pulse and ( B) response to a 26-s light pulse. The transient depolarizations ap- proach 0 mV in each case and is larger in (B) and reverses into a prolonged hyperpolarization. The potential re- turned to the resting level later. White light, 70 pW /cm2.

and growth retarding proteins (Ressho et al., 1994). For culture of other molluskan neurons (e.g., Helisomu and Lymnuea) conditioned me- dium, prepared by incubating CNS ganglia in the medium, is necessary for sprouting of neurons in vitro (Wong et al., 1983; Ridgway et al., 1991). Furthermore, experiments with murine nerve growth factor (NGF) and preabsorbtion of condi- tioned medium with anti-NGF suggest that a NGF- like molecule is produced by the molluskan ner- vous system ( Ridgway et al., 199 1 ).

As expected from Herman and Strumwasser's ( 1984) study of the localization of PNs, reduction of the eye to the basal retinal portion prior to disso- ciation enriched the cultures in monopolar neu- rons, which we have called PNs. We presume that the monopolar neurons are the PNs because their morphology best fits the characteristic of PNs in the intact retina and they may occur in clustersjust as they do in the outer layer of the intact retina. Earlier eye tissue reduction experiments suggested that PNs were responsible for generating the rhythm (Jacklet and Geronimo, 197 1 ) because the basal retina containing the majority of the PNs was necessary for generation of the circadian rhythm. Also, intracellular recording and dye marking of PNs (Jacklet et al., 1982) showed that PNs are the

source of the ON CAP activity, further supporting the idea that the circadian rhythm is likely gener- ated by the PNs.

Monopolar neurons identical in morphology to the PNs we studied here were previously dissoci- ated from the retina (Strumwasser et al., 1979), but successful recording of the membrane poten- tials and action potentials of these neurons was not reported. In retrospect, our success is probably due to our use of the whole cell patch recording tech- nique.

Our study establishes that the isolated monopo- lar neurons (PNs) are capable of producing action potentials in darkness. Shallow pacemaker poten-

,001 I Id

-100 L ' --- -

0 100 200. 30 time (rn4

400

B 500 T

lootL I

-40 40 -120 v iib

-200 1 Figure 11 Currents and current-voltage relationship taken from an R-type PR that had previously responded to light similar to Figure 10. ( A ) Currents recorded in response to voltage steps in the range from - 120 to +30 mV in 10-mV increments. Holding potential was -30 mV. Only late outward current is evident. Capacitance compensation failed to completely eliminate the capaci- tive currents. ( B ) The steady-state current-voltage rela- tionship derived from the cell in ( A ) . Currents were av- eraged over the last 2 ms of each step. Note, the current in ( B ) evoked in response to +40 mV is not represented in (A ) because it was off scale and required a gain change.

tials and slow irregular depolarizations lead to ac- tion potentials. Only small currents passing over the high membrane input resistance (5 GO) of the PNs are needed to produce the necessary depolar- ization. By Ohm’s law a current of 1 pA passing over a resistance of 5 GO is sufficient to produce a 5-mV depolarization. Such a small current ( 1 PA) can arise from the opening of a single ion channel (Hille, 1992).

Because recordings were only possible for less than 1 h we could not determine if a circadian rhythm of action potential frequency is produced by a single neuron, and we did not attempt to sam- ple the action potential frequency at all circadian phases because ofthe confounding problem of light exposure and other disturbances caused by record- ing. lntracellular recordings from isolated cultured BRNs of BulIa failed to show spontaneous impulse activity but did show a circadian rhythm of con- ductance change (Michel et al., 1993). It is likely that Bulla neurons are capable of producing a cir- cadian rhythm of impulse frequency under the proper recording and culture conditions, just as we expect Aplysia PNs can.

Isolated clusters of PNs in culture can produce very robust activity in darkness [see Figs. 5 (B) ,6] , apparently as a consequence of depolarizing pulsa- tile interactions among the PNs, which contribute to the depolarization leading to PN action poten- tials. The interactions seen in clusters are similar to those seen in the intact retina (Jacklet, 1969b; Jacklet et al., 1982), where the interactions are thought to be mediated by electrotonic coupling because blocking chemical synaptic activity does not prevent spontaneous or light evoked activity. Also, dye coupling occurs among PNs. These in- teractions likely occur over gap junctions, which have been found between PNs in ultrastructural studies ( Luborsky-Moore and Jacklet, 1977; Strumwasser et al., 1979). Electrical coupling, al- lowing dye coupling, among neurons is a promi- nent feature that contributes to CAP synchrony in the BRNs of Bulla also (Jacklet, 1988; Geusz and Block, 1992). Detailed studies employing simulta- neous recordings from several PNs in a cluster in vitro are needed to confirm the coupling mecha- nisms used in Aplysia PN interactions.

PNs attached to retinal fragments respond to light with sustained depolarization and a train of action potential of relatively constant frequency, just as PNs in the intact eye do (Jacklet, 1969b; Jacklet et al., 1982). The light induced depolariza- tion of PNs in the intact retina was thought to be mediated by electronic coupling to PRs, because

high magnesium artificial sea water did not block the light response (Jacklet et al., 1982). However, recently we observed that some isolated cultured PNs respond to light directly (Jacklet and Barnes, 1993b). Thus, the light induced depolarization of a PN attached to a retinal fragment may be direct, in part. Correlative support for the light induced depolarization of PNs is provided by evidence from Bulla that a reduced eye containing only BRNs, likely homologs of PNs, also appear to respond to light directly (Block and McMahon, 1984).

The PNs in the intact eyes are electrically cou- pled and fire in synchrony to produce CAP activity. In contrast, neurons in the suprachiasmatic nu- cleus (SCN), which is the major circadian clock nucleus in vertebrates, do not seem to be coupled. Recent evidence shows that isolated SCN neurons in culture produce independent circadian rhythms of impulse activity when the activity is recorded ex- tracellularly via electrodes implanted in the culture dish (Welch et al., 1995 ), and this activity does not seem to become synchronized. It will be interesting to make extracellular recordings from molluskan PN clusters, using electrodes implanted in the cul- ture dish, to see if they retain the ability to interact sufficiently to produce a circadian rhythm of syn- chronized action potential activity.

The ionic currents exhibited by a PN attached to a cluster comprise a set that is similar to the ionic currents recorded from an isolated PN. Those cur- rents are described elsewhere in detail (Jacklet and Barnes, 1993a; Jacklet and Barnes, 1995) and in- clude transient inward Na and Ca currents, a de- layed rectifier K current, a CI current, and an A current. The C1 current is induced by CI leaking into the cell, if KCI electrodes are used, and it is not likely to be important under physiological condi- tions. I t is seen in attached PNs also when KCI elec- trodes are used (Jacklet and Barnes, 1995 ). Inward currents in PNs in a cluster are likely carried by both Cat+ and Na+, but no attempt was made to separate them in this study. The delayed K current is modulated by the circadian clock because the current is larger at the circadian predawn phase than in the postdawn phase. Also, the delayed K current is enhanced by serotonin (Jacklet and Barnes, 1995 ), which is a neurotransmitter and po- tent phase-shifting agent of the Aplysia eye circa- dian rhythm (Corrent et al., 1978), and is localized in efferent nerve terminals in the eye (Jacklet, 1987).

In their detailed study Herman and Strumwas- ser ( 1984) described five morphological PR types in the retina: microvillus, tufted, whorled, flared,

Aplysia Retinal Pacemaker Neurons 27

and filiform. The largest, most conspicuous, and numerous cell is the microvillus PR, which corre- sponds to the R type (Jacklet, 1969b; Jacklet and Rolerson, 1982). It responds to light in the intact retina with a graded depolarization that reaches the 0-mV membrane potential level initially and then reverses to a prolonged hyperpolarization if the light pulse is long and intense ( Jacklet and Roler- son, 1982). Isolated R-type PRs identified by their morphology in this study responded in the same way, confirming that R-type PRs in culture retain their innate characteristics. Previous studies (Strumwasser et al., 1979) had shown that isolated R-type PRs in culture respond to light. Many of the other smaller PRs (i.e., tufted, whorled and flared) described by Herman and Strumwasser ( 1984) were seen in cell culture.

We demonstrated in this study that isolated PNs and clusters of PNs are capable of expressing the autorhythmic action potential activity (0.1-0.5 Hz) exhibited in recordings made from PNs in the intact eye (Jacklet, 1969b; Jacklet et al., 1982). Further refinement of the patch recording tech- nique in combination with the use of extracellular implanted electrodes in the culture dish, to record the PN activity over several days, should allow us to determine the membrane potential and current changes that occur over the course of several circa- dian cycles and allow us to understand the changes associated with the various phases of the circadian rhythm.

REFERENCES

BESSHO, C., ISHLHARA, H., KAWADA, H., YAMAURA, M., and SHIANO, S. ( 1994). Neurite outgrowth ofcul- tured Aplysia neurons and neurotrophic factors, Soc. Netirosci. Abst. 20:2 13.

BLOCK, G. and WALLACE, S. ( 1982). Localization of a circadian pacemaker in the eye of a mollusc. Bulla. Science 21 7: 1 5 5- 1 5 7.

BLOCK, G. D. and MCMAHON, D. G. (1984). Cellular analysis of the Bulla ocular circadian pacemaker sys- tem: 111. Localization of the circadian pacemaker sys- tem. J. Comp. Physiol. 155:365-378.

CORRENT, G., MCADOO, D., and ESKIN, A. (1978). Serotonin shifts the phase of the circadian rhythm from the Aplysia eye. Science 202:977-979.

GEUSZ, M. and BLOCK, G. (1992). Measurement of electrical coupling between circadian pacemaker cells ofthe Bullu eye. SOC. Neurosci. Abst. 18:l.

HERMAN, K. and STRUMWASSER, F. (1984). Regional specializations in the eye of Aplysiu, a neuronal circa- dian oscillator. J . Comp. Neurol. 230:593-6 13.

HILIE, B. ( 1992). Ionic Channels qj” Excitable Mem- branes, 2nd ed. Sinauer, Sunderland, MA.

JACKLET, J. ( 1969a). Circadian rhythm of optic nerve impulses recorded in darkness from isolated eye of Aplysia. Science 164:562-563.

JACKLET, J. ( I969b). Electrophysiological organization of the eye of Aplysia. J . Gen. Physiol. 53:2 1-42.

JACKLET, J. ( 1987). FMRF-amide-like immunoreactive efferent neurons responsible for CNS control of retinal circadian pacemaker neurons in the gasttropod, Bullu; comparison to the Aplysia system. In: Neurobiology, Molluscan Models. H. H. Boer, W. P. M. Geraerts, and J. Joosse, Eds. North Holland Publ., Amsterdam, pp.

JACKLET, J. ( 1988). Circadian pacemaker neurons: membranes and molecules. J . Physiol. (Paris) 83:

JACKLET, J. ( 1989). Cellular neuronal oscillators. In: Neuronal and Cellular Oscillators. J. Jacklet, Ed, Mar- cel Dekker, New York, pp. 482-527.

JACKLET, J. ( I99 1 ). Photoresponsiveness of Aplysia eye is modulated by the ocular circadian pacemaker and serotonin. Biol. Bull. 180:284-294.

JACKLET, J . and BARNES, S. ( 1993a). Two channel can- didates underlying circadian rhythmogenesis in iso- lated pacemaker neurons of Aplysia retina. Soc. Neu- rosci. Ahstr. 19:1702.

JACKLET, J. and BARNES, S. (1993b). Photoresponsive pacemaker neurons from dissociated retina ofdplysiu. NeuroReport 5:209-2 12.

JACKLET, J. and BARNES, S. (1995). Circadian phase differences in a retinal pacemaker neuron delayed rec- tifier K current and increased current induced by a phase-shifting neurotransmitter, serotonin. Physiolo-

JACKLET, J. and GERONIMO, J. (1971). Circadian rhythm: population of interacting neurons. Science 174:299-302.

JACKIET, J. and ROLERSON, C. ( 1982). Electrical activ- ity and structure of retinal cells of the Aplysia eye: 11. Photoreceptors. J. Exp. Biol. 99:38 1-395.

JACKLET, J., SCHUSTER, L., and ROLERSON, C. ( 1982). Electrical activity and structure of retinal cells of the Ap/.vsiu eye: I . Secondary neurones. J. Exp. Biol. 99:

KLEINFEID, D., PARSONS, T., RACCUIA-BEHLING, F., SLAZBERG, B., and OBAID, A. (1990). Foreign con- nections are formed in vitro by Aplysia cul~fornica in- terneurone LIO and its in vivo followers and non-fol- lowers. J . Exp. Bid. 154237-255.

LICKEY, M., WOZNIAK, J., BLOCK, G., HUDSON, D., and AUGTER, G. (1977). The consequences of eye re- moval for the circadian rhythm of behavioral activity in Aplysia. J. Comp. Physiol. 118: 12 1 - 143.

LUBORSKY-MOORE, J. and JACKLET, J. (1977). Ultra- structure of the secondary cells in the Aplysia eye. J. Ultrustritct. Rex 60:235-245.

MCMAHON, D., WALLACE, S., and BLOCK, G. (1984).

36-42.

164-17 1.

gist 38:A-22.

369-380.

28 Jackket et a/.

Cellular analysis of the Bulb ocular circadian pace- maker system: neurophysiological basis of circadian rhythmicity. J. Cornp Physiol. A 161:379-385.

MICHEL, S., GEUSZ, M., ZARITSKY, J., and BLOCK, G. ( 1993). Circadian rhythm in membrane conductance expressed in isolated neurons. Science 259:239-24 1.

OLSON, L. and JACKLET, J. ( 1985). The circadian pace- maker in the Apl-vsia eye sends axons throughout the central nervous system. J. Nezrrosci. 5 3 2 14-3227.

RIDGWAY, R., SYED, N., LUKOWIAK, K., and BULLOCH, A. (1991). Nerve growth factor (NGF) induces sprouting of specific neurons of the snail, Lymnaeu stagnalis. J. Neiirobrol. 22:377-390.

SCHACHER, S. and PROSHANSKY, E. ( 1983). Neurite re- generation by neurons in dissociated cell culture: mod- ulation by Aplysia hemolymph and the presence of the initial axonal segment. J. Neurosci. 3:2403-24 13.

STRACK, S. and JACKLET, J. (1993). Antiserum to an eye-specific protein identifies photoreceptor and pace-

maker neuron projections in Aplysia. J . Neurobiol. 24: 1-19.

STRUMWASSER, F., ALVAREZ, R., VIELE, D., and Woo- LUM, J. ( 1979). Structure and function of a circadian oscillator system. In: Biological Rhythms and Their Central Mechanisms. M. Suda, 0. Hayaishi, and H. Nakagawa, Eds. Elsevier/ North Holland Biomedical Press, Amsterdam.

SYED, N., BULLOCH, A,, and LUKOWIAK, K. (1990). In vitro reconstruction of the respiratory central pattern (CPG) of the mollusc Lyrnnaea. Science 250:282- 285.

WELCH, D., LOGOTHETIS, D., MEISTER, M., and REP- PERT, S. ( I995 ). Individual neurons dissociated from rat suprachiasmatic nucleus express independently phased circadian firing rhythms. Neuron 14:697-706.

WONG, R., MAR rEL, E., and KATER, S. ( 1983). Condi- tioning factors produced by several molluscan species promote neurite outgrowth in cell culture. J . Exp. Bid . 105389-393.