Odor Sensitivity of Cultured Lobster Olfactory Receptor Neurons is Not Dependent on Process Formation

The small size and thin, elongated morphology of olfactory receptor neurons (ORNs) has made the study of olfactory transduction difficult. The advent of patch-clamp recording ameliorated this situation and facilitated progress towards understanding olfactory transduction (reviews: Anholt, 1991; Firestein, 1991). Central to this effort has been the ability to dissociate ORNs from their surrounding epithelium in order to study them directly or in sustained primary culture. Dissociated ORNs are not only accessible for patching, they often assume a more compact form than their counterparts in situ that allows a reasonable space-clamp and facilitates diffusion of membrane-impermeant probes from the electrode to the site of transduction. While transduction is thought to occur in the cilia (outer dendrites, in invertebrates) of the ORNs (Kurahashi, 1989; Firestein et al. 1990; Lowe and Gold, 1991), at least some elements of the transduction cascade are not confined to the cilia. Specifically, cyclic-AMP-gated cation channels that are the effectors in the transduction cascade in amphibian ORNs also occur on the dendrite and soma of the cells, although in lower densities than on the cilia (Firestein et al. 1991; Zufall et al. 1991a). Indeed, this variability in density was exploited to obtain favorable channel density for recording (Firestein et al. 1991; Zufall et al. 1991a).

Previously, we reported preliminary evidence that cultured lobster ORNs respond to odors regardless of whether the cells had sprouted processes (Fadool et al. 1991). This observation raises the interesting possibility that, in vitro, all elements of the transduction cascade may be expressed and inserted into the soma of ORNs that fail to sprout processes. Given the ease of patching the soma compared to the extremely thin cilia (outer dendrites) and the ability to culture ORNs from vertebrates (Calof and Chikaraishi, 1991; Coon et al. 1989; Ronnett et al. 1991; Pixley and Pun, 1990) and other invertebrates (Stengl et al. 1989; Zufall et al. 1991b), such a phenomenon could be of general use for studying olfactory transduction.

Without the normal polarity of the cell, however, it must be established that applied ‘odors’ are activating what would otherwise be ciliary (dendritic) chemoreceptors. The need to establish the adequacy of odor stimuli is particularly important when studying ORNs from aquatic animals such as fish and lobsters. Adequate olfactory stimuli for many aquatic animals are the blood-borne components of prey, compounds such as amino acids, amines and nucleotides (review: Carr, 1990). Such compounds could be expected to activate cells as neurotransmitters or neuromodulators. In order to establish the usefulness of cultured lobster ORNs for analysis of transduction mechanisms, we have attempted to obtain functional evidence that cultured lobster ORNs with no processes or varying numbers of processes are morphs of the same type of cell and that the odorevoked properties of the cultured cells reflect those of lobster ORNs in situ.

The distinct clusters of the ORNs were dissected from the aesthetasc (olfactory) sensilla on the lateral antennular filament (olfactory organ) of adult specimens of the Caribbean spiny lobster, Panulirus argus Latreille. The clusters were enzymatically dissociated, and the resulting cells sustained in primary culture as described previously (Fadool et al. 1991). Briefly, the isolated clusters were incubated for 50min at 80revsmin⁻¹ on an orbital shaker in a 0.2 µm filter-sterilized solution of 2.5mg papain and 12mg L-cysteine in 10ml of Panulirus saline (PS) containing 1% penicillin, streptomycin sulfate and amphotericin B (Gibco). Proteolytic digestion was stopped by replacing the enzyme solution with low-glucose L-15 medium supplemented with L-glutamine, dextrose, fetal calf serum and BME vitamins. Cells were immediately plated on poly-D-lysine-coated glass coverslips. Cells were maintained at saturation humidity in a modular incubator chamber (Billups-Rothenberg) at 24°C. Neurite outgrowth in individual cells was recorded on a TL Panasonic 6050 time-lapse video cassette recorder. Images were later captured and subsequently analyzed using Image 1 analysis software.

Voltage- and odor-activated currents were recorded in the whole-cell configuration with an integrating patch-clamp amplifier (Dagan 3900). The analog signal was filtered at 5 kHz and digitally sampled every 4ms. Data acquisition and subsequent storage and analysis of the digitized records were carried out using pCLAMP software (Axon Instruments). Cells were viewed at 40X magnification with Hoffman optics. Patch electrodes, pulled from 1.8mm o.d. borosilicate glass, were fire-polished to a tip diameter of approximately 1.0µm (bubble number 4.8; Mittman et al. 1987). High-resistance seals (8.0–14 GD) were formed by applying gentle suction to the lumen of the pipette upon contact with the cell. In all experiments, cells were voltage-clamped at a holding membrane potential of −60mV. Membrane resistance changes were determined by injecting current sufficient to elicit 30mV, 300ms hyperpolarizing voltage steps into the cells from the holding potential. Each cell was photographed to allow correlation of soma size, length of process and cell morphology with physiological properties.

Odors were ‘spritzed’ on the cells for 120ms from a seven-barrel (six barrels were used) glass micropipette (Frederick Haer) coupled to a pressurized valve system (Picospritzer, General Valve). In most trials, fluorescein was used as an indicator in one barrel of the pipette, which was varied from one experiment to another, to position the tip of the pipette relative to the cell and to ensure that the delivered odorant completely surrounded the cell and its associated processes. The magnitude of the response to odors was found to be independent of which of the six barrels contained the odor. Dilution of the odor between the pipette and the cell surface, an average distance of two cell diameters, was estimated to be approximately 9%, based on the calculated K⁺ permeability method of Firestein and Werblin (1989). Odor concentrations are reported as the pipette concentration and are not corrected for this dilution.

The odors used were solutions of (1) an equimolar mixture that included (10⁻³ mol l⁻¹) betaine, glycine, lactic acid, taurine and trimethylamine oxide, referred to as S-1; (2) an aqueous extract of TetraMarin, a commercially available fish food, prepared as described earlier (Schmiedel-Jakob et al. 1990) and diluted 1000-fold, referred to as TET; and (3) single substances known to be effective odors for the lobster, which included (10⁻³ mol l⁻¹) adenosine monophosphate (AMP), arginine, ascorbic acid, betaine, cysteine, glycine, histamine, proline, taurine and trimethylamine oxide (TMAO). All odorant solutions were prepared daily in modified L15 medium and applied at the stated concentration unless otherwise noted.

Panulirus saline (PS) consisted of (in mmoll⁻¹) 458 NaCl, 13.4 KCl, 9.8 MgCl₂, 13.6 CaCl₂, 13.6 Na₂SO₄, 3 Hepes and 2 glucose; pH7.4. Modified L15 medium consisted of 50ml of Liebowitz L15 stock, 50ml of 1.6 times normal concentration of PS, 0.6g of dextrose, 0.026g of L-glutamine, and 0.01% gentamicin. The patch electrode solution consisted of (mmol l⁻¹) 30 NaCl, 11 EGTA, 10 Hepes, 1 CaCl₂, 180 potassium acetate and 696 glucose; pH7.0. All salts were obtained from Sigma.

The cultures consisted largely of the five morphological types of small (8–16 µm diameter soma) ‘neuron-like’ cells, described previously (Fadool et al. 1991): (1) soma only, (2) soma with bud, (3) unipolar, (4) bipolar or (5) multiprocess, four of which were used in the present study (Fig. 1A). The processes ranged from 3 to 160 µm long. Each of the four morphological types was present as early as 2h post-plating. Initially, the predominant form was ‘soma only’, but the proportion of each morphological type changed over time; the relative proportion of cells lacking processes decreased, with a concomitant increase in the proportion of cells with processes (Fig. 1B). To distinguish whether the change in the relative proportion of the morphological types reflected selective loss of cells lacking processes or process proliferation, or both, 218 ‘soma only’ cells were followed individually with digital time-lapse imaging for three consecutive days starting 2h after plating. 42 (20%) of the cells died within the observation period, indicating that a selective loss of ‘soma only’ contributed to the change in the relative proportion of the cell types. While many of the 176 cells that persisted failed to sprout processes, 61 of the cells sprouted processes throughout the observation period, becoming uni-, bi- and, eventually, multipolar. The latter finding supports our contention that the four morphological types of ‘neuron-like’ cells are morphs of a single type of cell.

That the four types of ‘neuron-like’ cells were morphs of a single type of cell was supported by the finding that 268 cells tested for their current–voltage relationship had similar voltage-activated properties. The total membrane current evoked in a typical cell by depolarizing voltage steps consisted of a transient inward current (≤570 µs duration) that activated around −30mV, followed by a much larger, prolonged outward current that activated around −20mV and persisted with little decay throughout the 15ms duration of the pulse (Fig. 2A). The magnitude of the inward and outward currents was independent of the number of processes on the cell (ANOVA, P=0.47 for inward currents and P=0.71 for outward currents) (Fig. 2B). A subsample of 12 cells, including at least one cell of each morphological type, had a mean input resistance (R_N) at rest of 1.1±0.2 GD and a membrane time constant (t₀) of 67.3±11.3ms. No detectable equalizing time constant could be measured in any of the morphs, including morphs with multiple processes.

472 cells (including the cells tested for their electrical properties) were tested for their ability to generate a current in response to stimulation with up to five different odors. The odor arrays usually, but not always, included the complex mixture, TET. 64% of the cells responded when they could be tested with at least one odor. This percentage increased to 89% when the cells could be tested with at least three different odors (N=182). Odors evoked a transient current that rose to a maximum over several hundred milliseconds and subsequently declined more slowly to rest (Fig. 3). The current could be of either polarity, depending on the cell and the odor tested and different odors could evoke currents of opposite polarity in the same cell (Fig. 3). In cells that could be tested with at least three different odors (N=182), the odors that were tested evoked only inward currents in 48 cells (26%), only outward currents in 58 cells (32%) and currents of both polarities in 56 (31%) cells. The remaining 11% of the cells did not respond to any of the odors tested.

Odor-evoked currents of both polarities were associated with an increase in membrane conductance, as indicated by a decrease in input resistance, when hyperpolarizing voltage steps were injected into the cells prior to and during odor-stimulation (Fig. 4). The input resistance (R_N) decreased significantly under odor stimulation from a mean of 1.1±0.1 to 0.7±0.1 GD for the inward current (N=8) and from a mean of 1.3±0.6 to 0.7±0.1 GD for the outward current (N=4) (paired t-test, P<0.05). Concomitantly, the membrane time constant (t₀), decreased significantly during odor stimulation from 59.6±12.6 to 21.6±4.1ms for the inward current (N=14) and from 82.5±20.7 to 34.6±9.5ms for the outward current (N=7) (paired t-test, P<0.05).

The latency to the onset of the odor-evoked currents, measured from the activation of the spritzer, ranged from less than 20ms to more than 1s, but typically was less than 100ms (Fig. 5). Overall, the mean latency to onset for the inward current, 186.2±31.3ms (N=100), was significantly longer than that for the outward current, 81.0±10.6ms (N=121) (t’ statistic, P<0.05). To determine whether this difference was caused by variations between cells, we performed a paired comparison of the latency in 19 cells that supported odor-evoked currents of both polarities. The mean latency in these cells for the inward current was not significantly longer than that for the outward current (paired t-test, P<0.05).

The peak amplitude of odor-evoked currents of both polarities increased with the concentration of the odor and saturated over 3–4 orders of magnitude (Fig. 6). The mean slope of the concentration–response function in the steepest region of the curve was 8.3±1.7pA per decade (N=3) for the inward current and 1.3±0.2pA per decade (N=7) for the outward current. Thresholds were less than 10⁻⁸ mol l⁻¹, the lowest concentration tested, for currents of either polarity.

The peak amplitude of currents of both polarities evoked by a single concentration (10⁻³ mol l⁻¹) of odor ranged, typically, from 5 to 85pA (Fig. 7A). The average magnitude of the inward current (39.2±3.0pA) was significantly greater than that of the outward current (20.1±1.4pA) measured across all cells and odors (N=386; t’ statistic, P≤0.05). The polarity and the magnitude of the currents evoked by odors were independent of the cell morphology for six different odors. The results for two of the six odors, proline ( N=111 cells) and taurine (N=102 cells), are shown in Fig. 7B. The peak amplitude of the odor-evoked currents of either polarity was also independent of the length of the process in cells bearing processes (N=55, correlation analysis, r>3.86) (Fig. 8A) and the diameter of the soma in cells lacking processes (N=60, correlation analysis, r>3.86) (Fig. 8B).

Single odors activated 14–77% of the cells (Table 1). The stimulatory effectiveness was betaine > histamine > glycine > proline > taurine > AMP > TMAO > ascorbate > arginine > cysteine. An equimolar mixture of five compounds (S-1: betaine, taurine, glycine, TMAO, lactate) ranked intermediate in stimulatory effectiveness, as did the complex mixture (TET), although the concentration of TET relative to that of the pure compounds is unknown. Culture medium itself, tested as a control, stimulated none of the cells tested, even though it contained some of the same amino acids as above as well as L-glutamine. Some odors selectively activated either inward or outward currents (Fig. 9). Taurine elicited inward currents twice as frequently as outward currents, while proline showed the opposite tendency. Betaine and ascorbic acid elicited strictly outward currents, although the number of cells tested in each of these instances was not large.

In a separate experiment, 53 cells that were determined to be odor-responsive by their ability to respond to TET were presented sequentially with the same five odors in order to determine their response spectra or ‘tuning’. The cells varied in the number of odors they responded to, as well as in the magnitude and the polarity of the response to any one odor (Fig. 10A). The mean breadth of responsiveness (H) for the sample population was 0.31±0.04 (Fig. 10B). Repeating this analysis using fewer odors or the same number of different odors yielded measures of H between 0.2 and 0.3 (data not shown).

The fact that the cells could be observed to sprout processes in culture, together with their common active electrical properties, strengthens our earlier contention that the various types of ‘neuron-like’ cells in the culture are morphs of one type of cell, the ORN (Fadool et al. 1991). This finding is consistent with the highly enriched source of the cultured cells; the somata were harvested from the lumen of the olfactory organ, which is literally filled with grape-like clusters of the somata of the ORNs (Grünert and Ache, 1988).

Although most neurons, as well as non-neuronal cells, are chemically irritable (Lerner et al. 1990), it is reasonable to interpret the responses of the cultured ORNs as olfactory. Most importantly, the cultured ORNs showed a diversity of response profiles (Fig. 10). Postsynaptic or neuromodulatory receptors, in contrast, would be expected to be much more homogeneously distributed across the population of cells and show either the same or, if possessing multiple receptors, a restricted set of response profiles across cells. For example, a histamine-gated chloride channel implicated in modulating the output of lobster ORNs in situ occurs on the soma of more than 95% of the cells (McClintock and Ache, 1989a). Second, the average breadth of responsiveness of the cultured ORNs (H=0.31) approximated that of lobster ORNs determined in situ with a similar, but not identical, array of compounds (H=0.2, Derby et al. 1984) and did not differ significantly from that of lobster ORNs determined in situ with the same array of compounds (H=0.35, Student’s t-test, P<0.05, pooled S.D.; M. Wachowiak, unpublished data). Third, as in their counterparts in situ, odors evoked both inward and outward currents in the cultured ORNs, and different compounds could evoke currents of opposite polarity in the same cell (McClintock and Ache, 1989b). Odors that tended frequently to elicit currents of one polarity in the cultured cells, e.g. proline and outward currents, also did so in lobster ORNs in situ (Schmiedel-Jakob et al. 1990; Michel et al. 1991). That currents of opposite polarity were associated with an increase in membrane permeability argues, as occurs in situ (Schmiedel-Jakob et al. 1990; Michel et al. 1991), that the odors were not up- and down-regulating a single conductance but, rather, were regulating two distinct conductances through two parallel transduction pathways. Finally, the threshold and dynamic range of the cultured ORNs matched those of lobster ORNs in situ in response to the same or similar types of odor molecules (Schmiedel-Jakob et al. 1989; W. C. Michel, unpublished data). We conclude that most and possibly all, of the chemosensitivity expressed by cultured lobster ORNs is sensory, i.e. is mediated by the same receptors and transduction pathways found in the outer dendrites of lobster ORNs in situ, and is not the result of activating synaptic, neuromodulatory or other receptors that may occur on the soma of lobster ORNs.

The inability of the culture medium to elicit responses from the cells when applied as a control stimulus, even though it contains amino acids shown to be odors (cysteine, glycine, arginine) and others (e.g. glutamine) that are potential odors, is not necessarily surprising. The cells were maintained and tested in the same medium and would be expected to adapt to continuous background stimulation, as do their counterparts in vivo (Borroni and Atema, 1988). However, adaptation to continuous background stimulation does shift the apparent threshold of the cells in vivo so that recording in medium might have lowered the overall sensitivity to those test stimuli present in the medium (cysteine, glycine, arginine). Indeed, the smallest percentage of cells tested responded to cysteine and arginine; in contrast, glycine was among the more effective odors tested (Table 1). The remaining test odors, including those used for threshold determinations, were not components of the culture medium.

The overall magnitudes of the current evoked by odors in cultured cells could differ appreciably from those in situ as a result of possible differences in the number of channels expressed in culture. It is interesting, therefore, that the average magnitudes of the odorevoked currents in the cultured ORNs (39.2±3.0pA, inward; 20.1±1.4pA outward) were within an order of magnitude of those evoked in lobster ORNs in situ. A 10-fold greater concentration of TET than that used in the present study elicited an average inward current of 25.4±5.1pA, while the same concentration of proline as that used in the present study elicited an average outward current of 4.5±0.6pA in lobster ORNs in situ (Michel and Ache, 1992). Odor-evoked currents of dendritic origin in situ would be electrotonically attenuated in the soma (where they were recorded), so the actual magnitude of the current in the dendrite would presumably be closer to that found in the cultured cells. Our results are consistent with the possibility that the cultured cells express a full complement of odor-activated channels.

The average amplitude of the odor-evoked inward currents in lobster ORNs is an order of magnitude smaller than those reported for salamander or rat ORNs (Firestein et al. 1991; Kurahashi, 1989; Lowe and Gold, 1991; Pixley and Pun, 1990), but is consistent with our still limited understanding of unitary currents that may underlie the macroscopic current. Inositol 1,4,5-trisphosphate, Ins(1,4,5)P₃, the suspected excitatory second messenger, activated 1–3 channels in most (49/56) cell-free patches taken from the plasma membrane of the cultured cells (Fadool and Ache, 1992). On the assumption that these Ins(1,4,5)P₃-activated channels are transductory, there are an estimated 330–1990 channels per cell, given a pipette tip diameter of 0.5–1.0 µm and a total cell surface area of 380 µm² (Fadool et al. 1991); or 0.9–5.3 channels µm⁻² membrane. Taking the smaller of the two Ins(1,4,5)P₃-activated unitary conductances (2pA, Fadool and Ache, 1992), a macroscopic current of average magnitude (39pA; Fig. 7) would require activation of as few as 20 channels, or only 1–6% of the available channels. Such a small fraction of recruited channels is consistent with the presumed sensory function of the cells (e.g. Firestein et al. 1991).

The ability of the cultured cells to respond to odors did not correlate with the number or even the presence of processes, implying that the receptor sites for odors were not necessarily confined to the processes. At least in vertebrate ORNs, odor receptors are thought to be selectively localized to the cilia. Deciliating the frog olfactory epithelium with Triton X-100, for example, completely abolishes the electro-olfactogram (Adamek et al. 1984). Lowe and Gold (1991) confirmed this hypothesis by focally stimulating the cilia of dissociated salamander ORNs. Whether somatic sensitivity to odors in cultured lobster ORNs is induced or is typical of mature lobster ORNs in situ is unclear. Previous attempts to record responses to odors in freshly isolated somata have been consistently unsuccessful (e.g. Anderson and Ache, 1985), suggesting that somatic sensitivity to odors in cultured ORNs may have been induced by placing the cells in culture. This conclusion would be consistent with the observation that molluscan neurons raised in suspension culture to prevent process formation insert postsynaptic receptors otherwise normally confined to processes in the soma (Wong et al. 1981). Nevertheless, Hatt (1990) found odor-gated channels on the soma of freshly dissociated crayfish chemosensory neurons that had the same structure–activity profiles and sensitivity as the intact cells. Ligandgated histamine receptors on the soma of lobster ORNs resist the same enzymatic protocol used in the present study (McClintock and Ache, 1989a). Therefore, we cannot eliminate the possibility that the receptors in question normally occur on the somata of mature lobster ORNs, but are destroyed by the enzymatic treatment required to dissociate the cells and fail to recover in the subsequent 4–6h over which in situ recordings are usually made. It is known that enzymatic digestion used to prepare cells for patchclamping alters neurotransmitter responsiveness in invertebrate neurons (Oyama et al. 1990) and eliminates odor-responsiveness in amphibian ORNs (Firestein and Werblin, 1989).

The latency to response was extremely brief, as short as 20ms, with the majority of the responses occurring within 125ms of valve (spritzer) activation. These values fall below the lower limit of the range of latencies reported for dissociated salamander ORNs, the other system for which comparable measurements are available (175–600ms; Firestein et al. 1990). Both systems lack a mucus barrier that would impede stimulus access to the receptor sites, so the latencies presumably reflect the actual time course of transduction. Breer et al. (1990) recently reported that levels of two second messengers in rat and insect ORNs, cyclic adenosine monophosphate and inositol 1,4,5-trisphosphate, respectively, peak in response to odor stimulation within 25–50ms. Thus, while the latency of the response of cultured lobster ORNs is considerably shorter than that of salamander ORNs, it is consistent with second-messenger-mediated transduction.

Thus, we show that lobster ORNs not only survive in culture but express the odor sensitivity and selectivity of their counterparts in situ, including the ability of odors to excite as well as to inhibit the cells. As the cultured cells are morphologically and, presumably, electrotonically (as shown by the absence of a measurable equalizing time constant) more compact than their counterparts in situ, often consisting of only a spherical soma, the ability to characterize the transduction pathways and analyze potential interactions between them should be greatly facilitated in vitro.

This investigation was supported by ONR N0014-90-J-1566, NIMH NRSA 1F31MH10124-01A1. The authors would like to thank Ms Leslie Van Ekeris for technical assistance, Ms Lynn Milstead and Mr James Netherton for assistance with the illustrations and Drs Paul J. Linser and James M. Fadool for technical discussions.