Phosphodiesterases (PDEs) are key regulators of cyclic adenosine monophosphate (cAMP), an important signaling molecule implicated in various physiological functions including addiction and memory [1,2,3]. We previously cloned long-, short- and supershort-isoform of Aplysia PDE4 (ApPDE4), which are mammalian PDE4D homologues [4,5]. Overexpression of the supershort ApPDE4 isoform impaired serotonin (5-HT)-induced synaptic facilitation in a depressed synapse [6], and knockdown of the ApPDE4 isoforms impairs 5-HT-induced synaptic facilitation [4,5]. Short- and long-form of ApPDE4 were localized to the plasma membrane and to both the plasma membrane and presynaptic terminals, respectively [4]. We also showed that the 20 N-terminal amino acids of the long-form are sufficient for membrane expression in HEK293T cells and targeting to the presynaptic terminals of Aplysia sensory neurons [4,7]. The targeting mediated by the 20 N-terminal amino acids may be due to direct association with the membrane as evidenced by the ability of the first 16 N-terminal amino acids of ApPDE4 long-form to associate with the membrane through hydrophobic interactions [7]. Additionally, the short-form was found to target the plasma membrane via nonspecific electrostatic interactions, including the membrane affinity of the hydrophobic N-terminal, basic domain, and the interaction between the upstream conserved region 1 (UCR1) and UCR2 domains, located between the extreme N-terminal and the catalytic region [7]. Moreover, oligomerization of the short- or the long-form, mediated through an interaction between the UCR1 and UCR2 domains, further enhanced targeting to the plasma membrane [7]. Thus, the different ApPDE4 isoforms localize to distinct cellular regions and regulate cAMP signaling differently.

Previously, using reverse transcription polymerase chain reaction (RT-PCR) with isoform-specific primer sets, mRNAs of ApPDE4 isoforms were found to be primarily expressed in various tissues including the central nervous system [4]. However, the detailed distribution of ApPDE4s mRNAs in Aplysia individual ganglion was unclear. In this study, we performed in situ hybridization (ISH) using a probe containing the PDE catalytic domain of ApPDE4 to clearly determine the distribution of all ApPDE4 mRNAs in Aplysia ganglions. Interestingly, the strongest ISH-positive signals in the symmetrical bag cell clusters of the abdominal ganglion. Mechanosensory neurons from the sensory cluster also were stained on the ventral surfaces of the right and left pleural ganglia. Our data support an evidence that ApPDE4 is expressed in mechanosensory neurons in Aplysia and play key roles in 5-HT-induce synaptic facilitation in Aplysia mechanosensory neurons.

In this study, we examined distributions of all PDE4 mRNA isoforms in Aplysia ganglions. Previously, we identified three ApPDE4 isoforms sharing the identical PDE catalytic activity domain [4,5,6,7], but it is not clear how many different isoforms exist in Aplysia. We thus chose a probe corresponding to the PDE catalytic domain, which is commonly found in all ApPDE4 isoforms. Digoxigenin-labeled ApPDE4 anti-sense RNA was used to detect the ApPDE4 isoforms. Its sense probe was used as a negative control and no signal was detected (data not shown). On the other hands, ISH signals were found in ganglionic nerve cells by ApPDE4 anti-sense probe (Fig. 1,2,3): the strongest signals in the abdominal ganglion (Fig. 1) and relatively weak signals in other ganglions (Fig. 2 and 3). Notably, symmetrical bag cell clusters were stained most strongly in the abdominal ganglion (Fig. 1-A1 and A2). Other types of neurons including R2, R14, L7, L2, and L11 were stained weakly but significantly on the dorsal surface of the abdominal ganglion. Additional ISH-positive neurons were detected in the medial aspect of the right and left rostral quarter ganglia. On the ventral surface of the abdominal ganglion, ISH-positive endocrine cells were dispersed around the brachial nerve, the vulvar nerve, and the R11 cell (Fig. 1-B1 and B2). Normally, bag cells have a high resting potential, and they do not show any spontaneous activity. If an electrical stimulus is applied to the pleuro-abdominal connectives (Pl-Ab in Fig 1-A2), the bag cells fire repeatedly for 30 min, performing what is called an "afterdischarge" [8]. During the afterdischarge, the level of cAMP inside bag cells increases [9], thereby leading to modulation of potassium channels [10,11] and subsequent stimulation of ELH release. Therefore, our data suggest that each ApPDE4 isoform play a different role in modulating ELH release via cAMP regulation, as similar with different roles of ApPDE4 isoforms in Aplysia sensory neurons [4,5,6]. Therefore, our data suggest that each ApPDE4 isoform plays a different role in modulating ELH release via cAMP regulation, as similar with different roles of ApPDE4 isoforms in Aplysia sensory neurons [4,5,6]. Future studies examining the regulatory roles of each ApPDE4 isoform in the afterdischarge in bag cells are warranted.

ISH-positive cells were also observed in cerebral, pleural and pedal ganglion. First, the dorsal and ventral surfaces of the cerebral ganglion contained clusters positive for ApPDE4s, but the signals were weaker than the bag cells' (Fig. 2). On the dorsal surface of the cerebral ganglion (Fig. 2A), A, B and G cluster cells were stained strongly and symmetrically. Interestingly, metacerebral (MCC) neurons were stained in the G cluster. C, D and F cluster cells were stained less than the other clusters, but the ISH-positive neurons were detected symmetrically on both sides. On the ventral surface of the cerebral ganglion, ISH-positive neurons were observed in the outer layers of the A, B, and G clusters. The serotonergic MCC neurons are involved in food-induced arousal by direct modulation of ingestion-related B21 neurons in the buccal ganglion [12]. The MCC neurons receive long-lasting excitatory synaptic input by nitric oxide (NO) and histamine which induce cAMP-mediated signaling pathways [13]. Therefore, ApPDE4 might play a role in food-induced arousal via indirect pathways such as cAMP-dependent signaling.

Second, we found ISH-positive neurons in each pleural ganglion (Fig. 3A-D). The giant cell, LP1 which was a homologous giant neuron of R2, was stained on the dorsal surface of the left pleural ganglion as similar intensity level with R2 (Fig. 3A and 2A). The LP1 and R2 are cholinergic neurons involved in mucus release from the body wall [14]. This suggests that ApPDE4 has a role in mucus release. Some ISH-positive cells were observed near the pleura-abdominal, cerebro-pleural and pleura-pedal connections of the right pleural ganglion. On the ventral surface of both the pleural ganglia, several mechanosensory neurons from the sensory cluster were stained (Fig. 3C and D). Last, ISH-positive cells were detected in the pedal ganglion and these also showed the symmetrical staining pattern (Fig. 3E-H). On the dorsal surface of both pedal ganglia, ISH-positive neurons were located near the base of P3 and the middle pedal nerve (P9) (Fig. 3E and F). Additional ISH-positive neurons were found on the ventral surface near the pedal-pleural connective, cerebral connective, and P1 nerve in a symmetrical pattern (Fig. 3G and H). Previous study using RT-PCR showed that the short- and long-form of ApPDE4 were strongly expressed in the pleural ganglion, but the supershort-form was weakly expressed [4]. In consistent with our previous results, we found the mRNA expressions of ApPDE isoforms in mechanosensory neurons of pleural ganglion (Fig. 3-C and D). These data suggest that short- and long-form of ApPDE4 in the pleural ganglion have main functions in 5-HT-induced synaptic plasticity in Aplysia.