Distinct mechanisms for synchronization and temporal patterning of odor- encoding neural assemblies

Science

Volumn 274, Issue 5289, 1996, Pages 976-979

  MacLeod, Katrina ^a   Laurent, Gilles ^a  

^a CALIFORNIA INSTITUTE OF TECHNOLOGY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

NEUROPHYSIOLOGY; ODORS; SENSORY PERCEPTION; SYNCHRONIZATION;

DESYNCHRONIZATION; IONOTROPIC AMINOBUTYRIC ACID; ODOR ENCODING NEURAL ASSEMBLIES; STIMULUS EVOKED OSCILLATORY SYNCHRONIZATION;

NEUROLOGY;

4 AMINOBUTYRIC ACID RECEPTOR;

ANIMAL TISSUE; ARTICLE; CONTROLLED STUDY; INHIBITORY POSTSYNAPTIC POTENTIAL; LOCUST; NERVE PROJECTION; NEUROPIL; NONHUMAN; ODOR; OLFACTORY SYSTEM; OSCILLATION; PRIORITY JOURNAL; SMELLING; SYNAPTIC INHIBITION;

ANIMALS; FEMALE; GABA ANTAGONISTS; GAMMA-AMINOBUTYRIC ACID; GRASSHOPPERS; MALE; MEMBRANE POTENTIALS; NEURONS, AFFERENT; ODORS; OLFACTORY PATHWAYS; PICROTOXIN; RECEPTORS, GABA; RECEPTORS, SENSORY; SENSE ORGANS; SYNAPTIC TRANSMISSION;

ANIMALIA; HEXAPODA; MAMMALIA;

EID: 0030575890     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.274.5289.976     Document Type: Article

References (47)

1. C. M. Gray, J. Comput Neurosci. 1, 11 (1994).
2. G. Laurent, Trends Neurosci., in press.
3. W. Singer and C. M. Gray, Annu. Rev. Neurosci. 18, 555 (1995);
4. A. K. Engel et al., Trends Neurosci. 15, 218 (1992).
5. J. G. R. Jefferys, R. D. Traub, M. A. Whittington, Trends Neurosci. 19, 202 (1996).
6. E. D. Adrian, J. Physiol. (London) 100, 459 (1942); W. J. Freeman, Science 133, 2058 (1961); J. Neurophysiol. 31, 349 (1968); M. Satou, Prog. Neurobiol. 34, 115 (1990); K. Mon and Y. Yashihara, ibid. 45, 585 (1995).
7. E. D. Adrian, J. Physiol. (London) 100, 459 (1942); W. J. Freeman, Science 133, 2058 (1961); J. Neurophysiol. 31, 349 (1968); M. Satou, Prog. Neurobiol. 34, 115 (1990); K. Mon and Y. Yashihara, ibid. 45, 585 (1995).
8. E. D. Adrian, J. Physiol. (London) 100, 459 (1942); W. J. Freeman, Science 133, 2058 (1961); J. Neurophysiol. 31, 349 (1968); M. Satou, Prog. Neurobiol. 34, 115 (1990); K. Mon and Y. Yashihara, ibid. 45, 585 (1995).
9. E. D. Adrian, J. Physiol. (London) 100, 459 (1942); W. J. Freeman, Science 133, 2058 (1961); J. Neurophysiol. 31, 349 (1968); M. Satou, Prog. Neurobiol. 34, 115 (1990); K. Mon and Y. Yashihara, ibid. 45, 585 (1995).
10. E. D. Adrian, J. Physiol. (London) 100, 459 (1942); W. J. Freeman, Science 133, 2058 (1961); J. Neurophysiol. 31, 349 (1968); M. Satou, Prog. Neurobiol. 34, 115 (1990); K. Mon and Y. Yashihara, ibid. 45, 585 (1995).
11. W. Rall et al., Exp. Neurol. 14, 44 (1966).
12. G. Laurent and M. Naraghi, J. Neurosci. 14, 2993 (1994); G. Laurent and H. Davidowitz, Science 265, 1872 (1994); G. Laurent, M. Wehr, H. Davidowitz, J. Neurosci. 16, 3837 (1996); G. Laurent, Neuron 16, 473 (1996);
13. G. Laurent and M. Naraghi, J. Neurosci. 14, 2993 (1994); G. Laurent and H. Davidowitz, Science 265, 1872 (1994); G. Laurent, M. Wehr, H. Davidowitz, J. Neurosci. 16, 3837 (1996); G. Laurent, Neuron 16, 473 (1996);
14. G. Laurent and M. Naraghi, J. Neurosci. 14, 2993 (1994); G. Laurent and H. Davidowitz, Science 265, 1872 (1994); G. Laurent, M. Wehr, H. Davidowitz, J. Neurosci. 16, 3837 (1996); G. Laurent, Neuron 16, 473 (1996);
15. G. Laurent and M. Naraghi, J. Neurosci. 14, 2993 (1994); G. Laurent and H. Davidowitz, Science 265, 1872 (1994); G. Laurent, M. Wehr, H. Davidowitz, J. Neurosci. 16, 3837 (1996); G. Laurent, Neuron 16, 473 (1996);
16. M. Wehr and G. Laurent, Nature, in press.
17. Seventy-seven adult, nonanesthetized locusts of both sexes were taken from a crowded colony and immobilized with one antenna intact as described (7). A window was cut open on the dorsal region of the head to gain access to the brain, which was subsequently desheathed after softening with crystals of protease type XIV (Sigma). The brain was superfused with locust saline at room temperature, Intracellular recordings were made from the soma or dendrites of local and projection neurons in the antennal lobe by means of glass micropipettes with a dc resistance of 80 to 150 megohm when filled with 3 or 0.5 M K acetate. Intracellular staining was carried out by iontophoretic injection of cobalt hexamine (6% aqueous solution), with 0.5-s, 0.5- to 2-nA pulses at 1 Hz. Histology and intensification were carried out as described [J. P. Bacon and J. S. Altman, Brain Res. 138, 359 (1978)]. Airborne odors (monomolecular and blends) were puffed onto one antenna as described (7). Local field potentials were recorded with saline-filled patch pipettes (1-to 2-μm tips) from the mushroom body calyx.
18. W. J. Freeman, in Olfaction, a Model System for Computational Neuroscience, J. L. Davis and H. Eichenbaum, Eds. (MIT Press, Cambridge, MA), pp. 225-249.
19. Physiological evidence that LN depolarization can lead to PN inhibition has been provided for Manduca sexta [T. A. Christensen et al., J. Comp. Physiol. 173A, 385 (1993)]. Evidence that transmitter release from LNs to PNs does not require sodium action potentials corroborates results in the turtle olfactory bulb, where depolarized mitral cells can inhibit themselves by a reciprocal synapse with granule cells, even in the presence of tetrodotoxin, a sodium channel blocker [ C. E. Jahr and R. A. Nicoll, J. Physiol. (London) 326, 213 (1982)].
20. Physiological evidence that LN depolarization can lead to PN inhibition has been provided for Manduca sexta [T. A. Christensen et al., J. Comp. Physiol. 173A, 385 (1993)]. Evidence that transmitter release from LNs to PNs does not require sodium action potentials corroborates results in the turtle olfactory bulb, where depolarized mitral cells can inhibit themselves by a reciprocal synapse with granule cells, even in the presence of tetrodotoxin, a sodium channel blocker [ C. E. Jahr and R. A. Nicoll, J. Physiol. (London) 326, 213 (1982)].
21. K. MacLeod and G. Laurent, data not shown.
22. B. Leitch and G. Laurent, J. Comp Neurol. 373, 487 (1996). We found direct synaptic contacts between GABA-reactive profiles (LN to LN), between non-GABA-react ve profiles (PN to PN), and between GABA-and non-GABA-reactive profiles (LN to PN and PN to LN), indicating that LNs also inhibit each other, that PNs likely excite each other, and that PNs excite LNs.
23. A small volums (∼1 pl, measured before injection into the brain by inaction in a drop of oil) of a selected drug was pressure-injected into the antennal lobe or mushroom body calyx with a picopump (43 experiments). The mechanical effect of injection was usually small, allowing maintained intracellular impalement of PNs. We checked that drug diffusion was restricted to the selected neuropils (probably because of efficient glial barriers surrounding the antennal lobe and mushroom body calyx) by using injections of identical volumes of cobalt hexamine. Drugs (diluted in saline) used were ionotropic GABA receptor blockers [picrotoxin (PCT, Sigma; final concentration, 0.1 to 10 mM, n = 32 experiments), gabazine (SR-95531, Research Biochemical Co.; 0.1 to 10mM, n = 6), and bicuculline methiodide (BMI, Sigma; 5 to 7.5 mM. n = 5)]. Of these, picrotoxin was the most effective, causing a complete block of oscillations at a bath concentration of 100 μM. Gabazine was less effective (complete block of oscillations at 5 mM). Bicuculline was totally ineffective, even at high concentrations (15). Other evidence suggests that histamine is an inhibitory neurotransmitter in the brain of crustaceans [B. J. Claiborne and A. I. Selverston, J. Neurosci. 4, 708 (1984); T. McClintock and B. W. Ache, Proc. Natl. Acad. Sci. U.S.A. 86, 8137 (1989)] and insects [ R. C. Hardie, Nature 339, 704 (1989)]. We thus tested the effectiveness of cimetidine (H2 receptor blocker, Sigma; 1 to 10 mM, n = 7) and pyrilamine (H1 receptor blocker, Sigma; 1 to 12 mM, n = 7) in blocking odor-evoked LFP oscillations. Although attenuating effects could be observed (cimetidine > pyrilamine), they always required much higher concentrations (several millimolar) than with picrotoxin. Because the insect histamine receptor channel is, like the insect ionotropic GABA receptor, a ligand-gated chloride channel, the effect of histamine receptor blockers might have been due to nonspecific binding to the GABA receptor channels. We therefore focus here only on the specific effects of picrotoxin.
24. A small volums (∼1 pl, measured before injection into the brain by inaction in a drop of oil) of a selected drug was pressure-injected into the antennal lobe or mushroom body calyx with a picopump (43 experiments). The mechanical effect of injection was usually small, allowing maintained intracellular impalement of PNs. We checked that drug diffusion was restricted to the selected neuropils (probably because of efficient glial barriers surrounding the antennal lobe and mushroom body calyx) by using injections of identical volumes of cobalt hexamine. Drugs (diluted in saline) used were ionotropic GABA receptor blockers [picrotoxin (PCT, Sigma; final concentration, 0.1 to 10 mM, n = 32 experiments), gabazine (SR-95531, Research Biochemical Co.; 0.1 to 10mM, n = 6), and bicuculline methiodide (BMI, Sigma; 5 to 7.5 mM. n = 5)]. Of these, picrotoxin was the most effective, causing a complete block of oscillations at a bath concentration of 100 μM. Gabazine was less effective (complete block of oscillations at 5 mM). Bicuculline was totally ineffective, even at high concentrations (15). Other evidence suggests that histamine is an inhibitory neurotransmitter in the brain of crustaceans [B. J. Claiborne and A. I. Selverston, J. Neurosci. 4, 708 (1984); T. McClintock and B. W. Ache, Proc. Natl. Acad. Sci. U.S.A. 86, 8137 (1989)] and insects [ R. C. Hardie, Nature 339, 704 (1989)]. We thus tested the effectiveness of cimetidine (H2 receptor blocker, Sigma; 1 to 10 mM, n = 7) and pyrilamine (H1 receptor blocker, Sigma; 1 to 12 mM, n = 7) in blocking odor-evoked LFP oscillations. Although attenuating effects could be observed (cimetidine > pyrilamine), they always required much higher concentrations (several millimolar) than with picrotoxin. Because the insect histamine receptor channel is, like the insect ionotropic GABA receptor, a ligand-gated chloride channel, the effect of histamine receptor blockers might have been due to nonspecific binding to the GABA receptor channels. We therefore focus here only on the specific effects of picrotoxin.
25. A small volums (∼1 pl, measured before injection into the brain by inaction in a drop of oil) of a selected drug was pressure-injected into the antennal lobe or mushroom body calyx with a picopump (43 experiments). The mechanical effect of injection was usually small, allowing maintained intracellular impalement of PNs. We checked that drug diffusion was restricted to the selected neuropils (probably because of efficient glial barriers surrounding the antennal lobe and mushroom body calyx) by using injections of identical volumes of cobalt hexamine. Drugs (diluted in saline) used were ionotropic GABA receptor blockers [picrotoxin (PCT, Sigma; final concentration, 0.1 to 10 mM, n = 32 experiments), gabazine (SR-95531, Research Biochemical Co.; 0.1 to 10mM, n = 6), and bicuculline methiodide (BMI, Sigma; 5 to 7.5 mM. n = 5)]. Of these, picrotoxin was the most effective, causing a complete block of oscillations at a bath concentration of 100 μM. Gabazine was less effective (complete block of oscillations at 5 mM). Bicuculline was totally ineffective, even at high concentrations (15). Other evidence suggests that histamine is an inhibitory neurotransmitter in the brain of crustaceans [B. J. Claiborne and A. I. Selverston, J. Neurosci. 4, 708 (1984); T. McClintock and B. W. Ache, Proc. Natl. Acad. Sci. U.S.A. 86, 8137 (1989)] and insects [ R. C. Hardie, Nature 339, 704 (1989)]. We thus tested the effectiveness of cimetidine (H2 receptor blocker, Sigma; 1 to 10 mM, n = 7) and pyrilamine (H1 receptor blocker, Sigma; 1 to 12 mM, n = 7) in blocking odor-evoked LFP oscillations. Although attenuating effects could be observed (cimetidine > pyrilamine), they always required much higher concentrations (several millimolar) than with picrotoxin. Because the insect histamine receptor channel is, like the insect ionotropic GABA receptor, a ligand-gated chloride channel, the effect of histamine receptor blockers might have been due to nonspecific binding to the GABA receptor channels. We therefore focus here only on the specific effects of picrotoxin.
26. Electrophysiological data (PN and LFP) were digitized post hoc from Digital Audio Tape with LabVIEW software and an NBMI016L interface (National Instruments, Austin. TX). Sliding cross-correlations were calculated piecewise over 200-ms-long windows shifted progressively in steps of 100 ms over each data set (simultaneous LFP and PN recordings) with the use of Matlab (The Math Works) as described (7). The cross-correlations in Fig. 2 are each constructed by averaging the cross-correlations calculated from each pair over n presentations of the odor (A: n = 20; B; n = 23; C: n = 16; D: n = 14; 1-s puffs at 0.1 Hz). The structure of these functions indicates the great consistency of the synchronization between PN and LFP (Fig. 2, A, C, and D). Because cross-correlations were performed on continuous data (LFP and membrane potential), the large variations of potential caused by PN action potential waveforms needed to be weighted down for pictorial representation. We thus artificially eliminated all PN action potentials before cross-correlation analysis (spikes clipped to ∼ 5 mV, traces band-pass filtered at 5 to 100 Hz with no phase shift), resulting in a continuous signal of amplitude, timing, and shape similar to the subthreshold synaptic waveform underlying each action potential. The periodic structure of cross-correlations calculated with the full-spike waveform was identical, except for the increased dynamic range of the cross-correlation function, rendering the pictorial display of small modulations in the function impossible (saturated at low or high amplitude).
27. A receptor antagonist, is also ineffective on most inhibitory synapses in insects but has some blocking effects on inhibition in the brain of the moth M. sexta [B. Waldrup, T. A. Christensen, J. G. Hildebrand, J. Comp. Physiol. 161A, 23, 1987]. BMI had no effect on the odor-evoked oscillations at concentrations as high as 7.5 mM.
28. Anti-GABA immunocytochemistry on the PN axon bundle that runs between antennal lobe and mushroom body revealed about 830 axons that were all negative for GABA (12). The inhibitory inputs received by KCs during odor responses (7) must therefore originate elsewhere. They could be from feedback inhibitory neurons activated by KCs or from feedforward inhibitory neurons excited by the PNs.
29. Although LFP spectra in these conditions were usually centered on the same peak frequency, they were often broader, showed a greater variance of peak frequencies from trial to trial, and contained substantial power at higher frequencies, including peaks at harmonics of the fundamental. The reason for this change in spectral composition probably resides in the different nature of LFP signals before and during PCT application to the mushroom body. Before PCT application, LFPs are close to sinusoidal and result mainly from synaptic currents caused in the dendrites of KCs, few of which produce action potentials (7). After PCT application, LFPs probably result mainly from coherent and suprathreshold activity of KCs (note the asymmetrical shape and the amplitude of these bursts) and thus contain active components caused by spike-generating currents.
30. F. W. Shürmann, Exp. Brain Res. 19, 406 (1974), The neurotransmitter or neurotransmitters expressed by KCs are presently unknown.
31. KCs do not express natural intrinsic oscillatory properties, as revealed by intracellular current injection in vivo (7), Because odors normally evoke suprathreshold responses in relatively few KCs and because the (intracellularly recorded) responses of KCs to an odor often comprise an early and maintained inhibitory component (7), odor-evoked KC inhibition likely results, at least in part, from a feedforward circuit.
32. Although olfactory bulb circuits appear to contain more neuronal subtypes than the insect antennal lobe, their fundamental design is markedly similar [J. G. Hildebrand, Proc. Natl. Acad. Sci. U.S.A. 92, 67 (1995); J. Boeck et al., in Chemosensory Information Processing, D. Schild, Ed. (Springer-Verlag, Berlin, 1990), pp. 201-227; G. M. Shepherd, Neuron 13, 771 (1994)].
33. Although olfactory bulb circuits appear to contain more neuronal subtypes than the insect antennal lobe, their fundamental design is markedly similar [J. G. Hildebrand, Proc. Natl. Acad. Sci. U.S.A. 92, 67 (1995); J. Boeck et al., in Chemosensory Information Processing, D. Schild, Ed. (Springer-Verlag, Berlin, 1990), pp. 201-227; G. M. Shepherd, Neuron 13, 771 (1994)].
34. Although olfactory bulb circuits appear to contain more neuronal subtypes than the insect antennal lobe, their fundamental design is markedly similar [J. G. Hildebrand, Proc. Natl. Acad. Sci. U.S.A. 92, 67 (1995); J. Boeck et al., in Chemosensory Information Processing, D. Schild, Ed. (Springer-Verlag, Berlin, 1990), pp. 201-227; G. M. Shepherd, Neuron 13, 771 (1994)].
35. W. W. Lytton and T. J. Sejnowski, J. Neurophysiol. 66, 1059 (1991); P. Bush and T. J. Sejnowski, Comput. Neurosci. 3, 91 (1996).
36. W. W. Lytton and T. J. Sejnowski, J. Neurophysiol. 66, 1059 (1991); P. Bush and T. J. Sejnowski, Comput. Neurosci. 3, 91 (1996).
37. M. A. Whittington, R. D. Traub, J. G. R. Jefferys, Nature 373, 612 (1995); S. R. Cobb et al., ibid. 378, 75 (1995).
38. M. A. Whittington, R. D. Traub, J. G. R. Jefferys, Nature 373, 612 (1995); S. R. Cobb et al., ibid. 378, 75 (1995).
39. M. Chaput and A. Holley, J. Physiol. (Paris) 76, 551 (1980); M. Meredith, J. Neurophysiol. 56, 572 (1986); A. R. Cinelli, K. A. Hamilton, J. S. Kauer, ibid. 73, 2053 (1995); M. Yokoi, K. Mori, S. Nakanishi, Proc. Natl. Acad. Sci. U.S.A. 92, 3371 (1995).
40. M. Chaput and A. Holley, J. Physiol. (Paris) 76, 551 (1980); M. Meredith, J. Neurophysiol. 56, 572 (1986); A. R. Cinelli, K. A. Hamilton, J. S. Kauer, ibid. 73, 2053 (1995); M. Yokoi, K. Mori, S. Nakanishi, Proc. Natl. Acad. Sci. U.S.A. 92, 3371 (1995).
41. M. Chaput and A. Holley, J. Physiol. (Paris) 76, 551 (1980); M. Meredith, J. Neurophysiol. 56, 572 (1986); A. R. Cinelli, K. A. Hamilton, J. S. Kauer, ibid. 73, 2053 (1995); M. Yokoi, K. Mori, S. Nakanishi, Proc. Natl. Acad. Sci. U.S.A. 92, 3371 (1995).
42. M. Chaput and A. Holley, J. Physiol. (Paris) 76, 551 (1980); M. Meredith, J. Neurophysiol. 56, 572 (1986); A. R. Cinelli, K. A. Hamilton, J. S. Kauer, ibid. 73, 2053 (1995); M. Yokoi, K. Mori, S. Nakanishi, Proc. Natl. Acad. Sci. U.S.A. 92, 3371 (1995).
43. K. L. Ketchum and L. B. Haberly, J. Neurosci. 13, 3980 (1993).
44. M. Wilson and J. M. Bower, J. Neurophysiol. 67, 981 (1992).
45. These mechanisms might use histamine or metabotropic GABA receptor channels, or both. They could also involve soluble gases such as nitric oxide, which can probably be produced by local neurons in insects [U. Müller and E. Buchner Naturwissenschaften 80, 524 (1993)] and the terrestrial mollusc Umax maximus [ A. Gelperin, Nature 369, 61 (1994)].
46. These mechanisms might use histamine or metabotropic GABA receptor channels, or both. They could also involve soluble gases such as nitric oxide, which can probably be produced by local neurons in insects [U. Müller and E. Buchner Naturwissenschaften 80, 524 (1993)] and the terrestrial mollusc Umax maximus [ A. Gelperin, Nature 369, 61 (1994)].
47. Supported by a NSF grant and Presidential Faculty Fellow Award to G.L. We thank E. M. Schuman, N. Kopell, and members of the Laurent lab for helpful discussions and comments.