the 2nd roots, while the endings within the neuropil of
the ganglion appear to in¯uence the central circuitries
concerned with motor programs (see below).
Intrinsic physiological properties
Of the two pairs of 5HT-containing neurosecretory
neurons, the best studied are the A1 pair (A1-5HT cells).
Although occasionally silent, these cells usually are
spontaneously active in the range 0.5±3 Hz. Recordings
from cell bodies show large overshooting action poten-
tials with prominent after hyperpolarizations, both of
which are typical of invertebrate neurosecretory neurons
(Beltz and Kravitz 1987; Ma et al. 1992). Spontaneous
®ring continues in the absence of calcium or with cobalt
added to the bathing medium, indicating that the action
potentials are not synaptically driven, although their size
and shape are altered under these conditions. In the
presence of 100 nmol l
)1
tetrodotoxin, action potentials
are completely abolished leaving no residual oscillations,
suggesting that a sodium current underlies the sponta-
neous activity (Cromarty et al. 1999). The after-hyper-
polarization is reduced or eliminated by superfusion
of preparations with tetraethyl ammonium chloride
(0.5±2 mmol l
)1
), 4-aminopyridine (100 nmol l
)1
) or
charybdotoxin (10 nmol l
)1
) (Cromarty et al. 1999).
These and other pharmacological studies suggest that
calcium-activated BK channels are important contribu-
tors to the after-hyperpolarization. Molecular studies
demonstrate that all isoforms of the shab form of the
shaker family of potassium channels are missing in the
A1-5HT cells, or are present at levels below the limit of
detection of our methods, but are present in all other
neuron types examined so far (Schneider et al. 1999).
Autoinhibition
A particularly interesting property of the A1-5HT neu-
rons is that they show a pause in their ®ring after a
period of high-frequency activation triggered by injec-
tion of current through an intracellular recording elec-
trode (Fig. 2) (Heinrich et al. 1999). We call this pause
``autoinhibition'' and it resembles the ``postactivation
inhibition'' seen in vertebrate serotonergic neurons from
the midline raphe nuclei (Aghajanian and Van-
derMaelen 1982). The prevailing explanation of the au-
toinhibition in vertebrate neurons is that it is due to
released 5HT acting back on 5HT
1a
receptors located on
the somata, axons and dendrites of the amine neurons
(for review see Aghajanian et al. 1990). Although the
autoinhibition seen in lobster neurons resembles that
seen in the vertebrate cells, the mechanism is dierent in
that it appears to be an intrinsic property of the cells. We
believe this to be the case because: (1) we still see auto-
inhibition in saline with no added calcium or with cobalt
added, when we see no remaining synaptic activity; and
(2) we still see the inhibition in nerve cords from animals
depleted of 5HT through use of the drug 5,7 dihydrox-
ytryptamine (5,7-DHT). The duration of the autoinhi-
bition is directly related to the magnitude and duration
of the period of high frequency stimulation, but is inv-
ersely related to the initial ®ring rate of cells over their
normal range of ®ring (0.5±3 Hz). If the tonic release of
5HT from cells is a key part of how amine neurons work
(see below), such a mechanism could serve to maintain
uninterrupted elevated levels of 5HT in target areas
when cells are ®ring at the higher rates.
Fig. 1 Reconstruction of 1st abdominal serotonin-containing (A1-
5HT) cell from intracellular injection of the enzyme horseradish
peroxidase. After physiological identi®cation of the A1-5HT cell, it
was injected with the enzyme horseradish peroxidase which was
allowed to diuse for 12±72 h (details in Beltz and Kravitz 1987).
Tissues then were ®xed and a reaction product was generated using
diaminobenzidene as substrate. The cell morphology was traced using
a computer reconstruction system. The drawing is a composite of two
separate injections, one of which showed better morphology in ganglia
A1 through T4, the other of which showed better morphology in T2
and T3. The inset diagram (left side of ®gure) is an artist's
reconstruction of a typical A1-5HT cell. The A1-5HT neurons have
two sets of ending in every anterior ganglion through the sub-
esophageal: one is in the central neuropil regions of the ganglion, the
other is along the second thoracic roots in peripheral neurosecretory
regions. See text for further details. This ®gure is slightly modi®ed
from Fig. 3A and Fig. 5 of Beltz and Kravitz (1987)
225
Synaptic inhibition and excitation
Pharmacological responsiveness
The rate of spontaneous ®ring of A1-5HT neurons is
reduced by bath application of either octopamine or
c-aminobutyric acid (GABA), and is increased initially,
followed by a prolonged reduction, after bath applica-
tion of 5HT (Ma and Weiger 1993; Heinrich et al. 1999).
Proctolin, which co-localizes with 5HT in these neurons,
increases the ®ring of the cells. Thus two modulators
(5HT and proctolin), which are co-localized in the same
cell (Siwicki et al. 1987), have predominantly opposing
physiological eects. Since the proportions of amine and
peptide released at dierent frequencies of stimulation
should vary (with more peptide released at the higher
frequencies of ®ring), the physiological consequences of
stimulating the A1 neurons also should vary, depending
on the ®ring frequency.
Inhibition
The 5HT-containing neurosecretory neurons receive
a constant barrage of spontaneous inhibitory input at
a frequency of around 3±5 Hz (Ma et al. 1992; Weiger
and Ma 1993). The inhibitory post-synaptic potentials
(IPSPs) are synchronized among the T5 and A1 ganglion
5HT cell pairs, and fall into three distinct size categories.
The most common of these are small (0.4±1.5 mV) and
originate from putative GABAergic neurons in the 3rd
abdominal ganglion. Inhibition arising from this source
is blocked by picrotoxin and eliminated by cutting or
blocking the connectives between the 2nd and 3rd
abdominal ganglia. While this suggests that spontane-
ously active GABAergic neurons in the A3 ganglion are
the source of the IPSPs, picrotoxin is not a completely
selective blocker of GABAergic input in crustaceans
(Marder and Paupardin-Tritsch 1978, 1980; Lingle and
Marder 1981). Upon eliminating this input to the A1
cells, the spontaneous ®ring rates of 5HT cells are in-
creased by about 50%, suggesting that these cells are
under constant inhibitory regulation (Weiger and Ma
1993). Large, slow IPSPs can be triggered in the A1 cells
with connective stimulation (HoÈrner et al. 1997; Hein-
rich et al., submitted). These seem to arise from ®bers
that traverse the entire length of the lobster ventral nerve
cord. A particularly interesting aspect of these slow in-
hibitory responses is that they are simultaneous with the
appearance of EPSPs in the octopamine-containing ne-
urosecretory neurons (R. Heinrich et al., unpublished
observations). Whatever the source of this input, it may
be an important part of the machinery involved in
governing opposing actions of the two amines. More-
over, unilateral stimulation of either anterior or poste-
rior connectives, leads to the appearance of bilateral
inhibitory synaptic responses in A1-5HT neurons. The
slow IPSPs are abolished after high frequency ®ring of
the A1-5HT cells, under conditions that produce auto-
inhibition in these cells, and recover back to their
original size over the next several minutes. Fast inhibi-
tory synaptic responses do not appear to be blocked by
the high frequency pre-®ring.
Excitation
Spontaneous
excitatory
post-synaptic
potentials
(EPSPs) also are seen in recordings from the A1-5HT
neurons, but these are best seen after blocking IPSPs
Fig. 2 Autoinhibition of A1-
5HT neurons. An intracellular
recording from a spontaneously
active 5HT neuron is shown on
the lower part of the ®gure.
With high frequency ®ring of
the cell through the intracellular
electrode there is a pause in the
®ring of the cell (autoinhibi-
tion). The duration of the au-
toinhibition period is directly
related to the magnitude of the
stimulation, but is inversely
related to the initial spontane-
ous ®ring rate (inset diagram).
See text for details. The inset
diagram is reprinted from
Fig. 4 of Heinrich et al. (1999)
226