The effect of serotonin (10^-9 M) exposure in vivo on the rate of 5-bromo-2-deoxyuridine (BrdU) incorporation into the neuronal precursor cell lineage was assessed (Figure 2). Serotonin had no impact on the number of BrdU-labeled cells in the niche (first-generation precursors) or migratory streams (second-generation precursors) relative to unexposed control animals (Figure 2A). However, serotonin treatment did increase the numbers of BrdU-labeled cells in the LPZ of cluster 10 (Figure 2B). These data suggest that serotonin does not influence all cells in the neuronal ancestral lineage, but rather that it alters the cell cycle only in the proliferation zones, which are composed of late second-generation precursors and their progeny (Figure 1D).

When total RNA sampled from the brains of crayfish P. clarkii was analyzed with RT-PCR, a 515-bp product for 5-HT_1α (Figure 3, lane 2) and a 546-bp product for 5-HT_2β (Figure 3, lane 3) were revealed. As an internal control, PCR for 18S rRNA from the same cDNA template was also performed, and a 427 bp product was detected (Figure 3, lane 4). Nucleotide sequence analysis verified that the PCR products correspond to the targeted cDNA fragment of P. clarkii 5-HT_1α [GenBank:EU131667] and 5-HT_2β [GenBank:EU131666].

In situ hybridization with the antisense riboprobe revealed the expression of 5-HT_1α mRNA in several neuronal cell clusters, but labeling was not detected in fiber tracts or synaptic neuropil regions (Figure 4). Nonspecific hybridization signal was not found, as shown in a comparable brain tested with the sense riboprobe (Figure 4I). In the protocerebrum, cells of cluster 6, which lies at the anterior edge of the median protocerebrum and extends through the brain from dorsal (Figure 4B) to ventral (Figure 4A), contained strong labeling for the 5-HT_1α transcript (Figure 4C), as did cells in clusters 7 and 8 (Figure 4A). In the deutocerebrum, prominent hybridization signals were found in cells of clusters 9 (Figure 4A, D) and 10 (Figure 4E), the latter of which can be seen in both ventral (Figure 4A) and dorsal (Figure 4B) views of the brain. In some preparations, a region of reduced labeling was found in the middle of cluster 10 (Figure 4E, asterisk). In situ hybridization in combination with immunostaining for glutamine synthetase (GS) (Figure 4H, arrowheads), which labels the niche and migratory streams 10 , revealed in some preparations that there is a lower 5-HT_1α expression region (arrow in Figure 4H) at the point where the GS-labeled migratory stream inserts into cluster 10. In addition, cells in deutocerebral clusters 11 (Figure 4B), 12 and 13 (Figure 4A) label with the 5-HT_1α receptor transcript with varying intensities. In the tritocerebrum, neuronal somata in lateral cluster 16 (Figure 4A, F) and medial cluster 17 (Figure 4A, G) are distinctly labeled.

In situ hybridization with the antisense riboprobe for 5-HT_2β mRNA (Figure 5) revealed an expression pattern similar to 5-HT_1α mRNA, in that the signal is seen only in cell bodies and not in neuronal fibers or neuropil regions. Overall, labeling for 5-HT_1α receptor is more prevalent than for 5-HT_2β receptor, as many cell clusters display 5-HT_1α but not 5-HT_2β; this result could be due to differing sensitivities of the probes or to actual differences in abundance of message. The 5-HT_2β transcript was found mainly in neuronal clusters 9 (Figure 5A, C) and 10 (Figure 5A, B, D). Two pairs of cells in cluster 11 in the medial deutocerebrum were also labeled (Figure 5B, E). In the ventromedial tritocerebrum, four to six cells in cluster 17 consistently expressed the 5-HT_2β transcript (Figure 5F). Unlike 5-HT_1α labeling (Figure 4E, H), in situ hybridization for 5-HT_2β did not reveal a lower expression area in the LPZ of cluster 10. Correspondingly, the insertion point of the glutamine synthetase-positive migratory stream in cluster 10 is intensely labeled for the 5-HT_2β transcript, similar to the other regions of cluster 10 (Figure 5D, G, H). The sense riboprobe showed no nonspecific labeling (Figure 5I).

To define the distribution of the 5-HT_1α receptor, single and double antibody staining methods were used (Figure 6). In the protocerebrum, strong 5-HT_1crust-immunopositive terminals are found in lateral extensions of the protocerebral bridge (Figure 6J-Ji). Posterior to the protocerebral bridge, in the anterior and posterior median protocerebral neuropils (AMPNs and PMPNs, respectively), punctate labeling is also found (Figure 6B). Particularly intense staining is found in neuronal terminals that compose conical and spherical glomeruli in the olfactory (OL) (Figure 6I, Ii) and accessory (AL) (Figure 6A, E) lobes, respectively. Strong cytoplasmic staining for 5-HT_1α is also found in some neuronal somata in several different cell clusters where homogeneous punctate immunoreactivity fills the cytoplasm, leaving the nucleus negatively stained in the middle of the soma: clusters 6 (Figure 6C), 9 (Figure 6G-Gi), 10 (Figure 6F), 11 (Figure 6K), 16 (Figure 6D-Di) and 17 (not shown). Because cytoplasmic labeling in cells is so intense, it is difficult to distinguish membrane-associated labeling; the presence of such distinct cytoplasmic labeling, however, may suggest the presence of newly synthesized or recycled receptor (but see also Discussion).

In situ hybridization for the 5-HT_1α transcript revealed a region of reduced expression in cell cluster 10. In some (but not all) preparations, the area of lower 5-HT_1α expression was superficial and easily seen right after hybridization (Figure 4E, H). In order to precisely localize this region, coronal serial sections (20 μm in thickness) were cut from the post-hybridized brains. Indeed, in the middle of the densely packed cells in cluster 10, some cells were found to express a reduced, but clearly detectable (Figure 7A, inset), level of 5-HT_1α mRNA in comparison with the surrounding cells (Figure 7A-C, sections from ventral to dorsal). To explore this observation further, in situ hybridization for 5-HT_1α receptor mRNA was combined with BrdU immunofluorescent labeling to reveal the locations of proliferating cells. In some cases, the BrdU-labeled cells (Figure 7E, blue) were positioned in the middle of the reduced 5-HT_1α expression region (Figure 7D, F, red) in cluster 10.

Immunocytochemical techniques were also used to ask whether 5-HT_1α receptor protein is found in neuronal precursor cells in the niche, migratory streams or proliferation zones of clusters 9 and 10. GS labeling reveals the niche (Figure 8A) and streams (Figure 8B); BrdU labeling identifies S phase cells in the streams and proliferation zones (Figure 8C-G). Precursor cells in the neurogenic niche (Figure 8A) and their daughters in the proximal and medial parts of the streams (Figure 8B) do not label for 5-HT_1α receptor. However, in the distal parts of the streams near the proliferation zones (Figure 8E) and in the proliferation zones (Figure 8F, G), some BrdU-labeled cells show distinct receptor labeling. Occasionally, BrdU-positive cells in the LPZ and MPZ are also seen that are not immunoreactive for the 5-HT_1α receptor (Figure 8C, D). This variability in receptor labeling among neuronal precursors in the proliferation zones suggests either that serotonin may regulate only a subset of the second/third-generation precursor cell population or that serotonin receptors other than 5-HT_1α may be present on these cells.

5-HT_2βCrust immunoreactivity is found in somata in several cell clusters in the crayfish brain (Figure 9), although in most cells this tends to be weaker than labeling observed for the 5-HT_1α receptor (Figure 6). This could be due to differences in either the affinities of the two antibodies or in the abundance of these two receptors. Similar differential labeling was also observed for 5-HT_1α and 5-HT_2β mRNA distribution. Notably, the medial giant cells (Figure 9A) and their long processes (Figure 9B, arrows) projecting toward cluster 6 label intensely with 5-HT_2βCrust. Cells in clusters 6, 9, 10, 16 and 17 are also labeled. The cells in the LPZ and MPZ are more intensely labeled than other cells in their respective cell clusters (LPZ, cluster 10; MPZ, cluster 9). While immunoreactivity of cells in these regions is consistent, it is relatively weak. Likewise, the olfactory and accessory lobes, which contain projections of somata in clusters 9 and 10, are distinctly, but weakly, labeled (not shown).

Double immunofluorescence for 5-HT_2βCrust and glutamine synthetase or BrdU revealed the presence of homogeneous cytoplasmic labeling for 5-HT_2β protein in S phase cells in the distal migratory streams near the proliferation zones (Figure 10C, Ci, E, Ei), as well as in the proliferation zones in clusters 9 and 10 (Figure 10D, Di, F, Fi). However, staining in the LPZ was more intense for 5-HT_2β than for 5-HT_1α. In contrast, cells in the neurogenic niche (Figure 10A) and proximal and medial parts of the migratory streams (Figure 10B) do not label for this receptor. The distribution of the 5-HT_2β protein in the precursor cells is therefore very similar to that observed for the 5-HT_1α receptor (see above and Figure 8), although the relative labeling intensities for these receptors are sometimes different. For instance, immunocytochemical staining for 5-HT_2β is stronger in the LPZ than in cluster 10 cells outside the proliferation zone, while the LPZ is more weakly labeled for 5-HT_1α mRNA than surrounding cluster 10 cells.

In order to understand which type of serotonin receptors are involved in the regulation of cell proliferation and neurogenesis after serotonin treatment, the 5-HT_1α agonist quipazine maleate salt (QMS) was administered. This agonist was chosen because it has been shown to have the highest potency and efficacy in activating 5-HT_1α receptors from P. clarkii compared with nine other pharmacological agents, and because it shows no activity with P. clarkii 5-HT_2β receptors 20 . The proliferating cells in cluster 10 were localized using BrdU detection of the nuclei of S phase cells. A gradual increase in the number of BrdU-labeled cells in the LPZ of cluster 10 was observed with increasing QMS concentrations from 2.2 × 10^-11 M to 2.2 × 10^-9 M (Figure 11A). The activation of 5-HT_1α receptors with a 10-hour exposure to 2.2 × 10^-9 M of QMS induced a significant increase (38%; P < 0.05) in the number of BrdU cells relative to controls (Figure 11A). Higher concentrations of QMS did not alter the number of BrdU-labeled cells in cluster 10, a finding that is in agreement with our previous dose-response analysis for serotonin in lobster brains 15 .

Experimental animals were treated with different concentrations of the 5-HT_2β antagonist methiothepin mesylate salt (MMS). MMS was chosen because prior studies 20 demonstrated that this antagonist had the highest efficacy on P. clarkii 5-HT_2β receptors among 29 antagonists tested, and no detectable activity on P. clarkii 5-HT_1α receptors. In the LPZ in cluster 10, a gradual decrease in BrdU-positive cells was observed with increasing MMS concentrations from 10^-10 M to 10^-5 M versus controls (Figure 11C). Blocking of 5-HT_2β receptors for 10 hours with MMS at 10^-8 to 10^-5 M caused a significant decrease (25%; P < 0.05) in BrdU-labeled cells over control levels.

In order to know whether 5-HT_1α and 5-HT_2β receptors also influence the numbers of proliferating cells in the niche and migratory streams, we counted the number of BrdU-positive cells in these regions after receptor agonist and antagonist treatment of live animals. In contrast to our findings in the LPZs in cluster 10, neither the 5-HT_1α agonist QMS (2.2 × 10^-9 M) nor the 5-HT_2β antagonist MMS (10^-8 M) caused significant changes in the numbers of BrdU-labeled cells in the niche or stream versus control levels (Figure 11B, D). There is some variability in the baseline number of BrdU-labeled cells in the QMS and MMS treatment groups; these range from one to four BrdU-labeled cells in the niche, and three to ten cells in the streams. We know that the rate of neurogenesis decreases as the animals grow and age 25 , and the variability in baseline counts is attributed to the range of animal sizes used. The size of control and experimental animals within each treatment group were, however, tightly controlled. The number of M-phase cells labeled with an antibody against phospho-histone H3 confirms the effects of QMS and MMS on the numbers of dividing cells in the niche, migratory streams and cell cluster 10 (data not shown).

The distribution of serotonin transporter (SERT) also was assessed immunocytochemically (Figure 12). Anti-SERT immunoreactivity is found widely in cells in clusters 6, 9, 10, 11, 16 and 17 (Figure 12A, C, D). Labeling in cluster 10 is particularly intense and appears to be localized to every cell in the cluster (Figure 12C, D), which is composed of 'globuli' cells whose somata are characteristically small and distinct from other brain neurons in that the cell nuclei virtually fill the cell bodies (Figure 12D) 26 . However, as with the localization of 5-HT_1α mRNA, the LPZ in cluster 10 is more lightly labeled than the surrounding cells (Figure 12A). The medial giant cells, which labeled strongly for 5-HT_2β receptor, also bind the anti-SERT antibody (Figure 12B). Of particular interest for the present study, the neurogenic system, including the niche cells and their fibers that compose the streams, is intensely labeled (Figure 12E, F).

There are similarities and some differences between the immunocytochemical labeling for the 5-HT_1crust antibody presented here, and the prior findings of Spitzer et al. 24 using the same antibody and crayfish species. While labeling is found in the same cell clusters and neuropils as in the earlier study, in our experiments neuropil regions (for example, olfactory and accessory lobes) were most intensely labeled, while in the prior studies cell clusters were most intensely labeled. We also found that immunoreactivity in cell clusters and neuropils was consistent, in contrast to the findings of Spitzer et al. 24 , where the intensity of labeling with the 5-HT_1crust antibody, as well as the numbers of cells in the various cell clusters, was highly variable between animals. It was suggested that the variability observed in the prior studies might be 'indicative of the physiological state of the animal' 24 . In this regard the animals used in our assays, for unknown reasons, may have been more uniform. It is more likely, however, that these differences are due to modifications in fixation and processing methods. Spitzer et al. also observed an uneven distribution of labeling in the cytoplasm and around the periphery of irregularly shaped somata in the brain; cells showing receptor labeling in our studies were spherical or ovoid, and labeling was homogeneously distributed throughout the cytoplasm. Possible membrane-associated labeling was observed in both studies, a point that requires further investigation at the ultrastructural level.

The immunocytochemical localization of 5-HT_1α and 5-HT_2β receptor protein in the present study identifies sites of serotonin's action in the brain of the crayfish P. clarkii. In addition, the in situ hybridization results establish which cells synthesize 5-HT receptor proteins and contribute these to the neuropils and to the system generating new neurons in the adult brain. However, the most intense receptor labeling was found in neuropil regions, while 5-HT_1α and 5-HT_2β mRNA was predominantly localized to cell somata. For example, we observed intense labeling for the 5-HT_1α protein, as well as distinct but weak labeling for the 5-HT_2β receptor, in the olfactory and accessory lobes. These neuropil regions contain axonal fibers and synapses of local and projection interneurons in cell clusters 9 and 10, respectively. However, the mRNA for both 5-HT receptors is primarily found in the somata of cells in clusters 9 and 10 and not in the neuropils.

There are two possible explanations for the distinct intracellular distributions of receptor mRNA and protein. Firstly, this may suggest that there is little or no synthesis of the serotonin receptor protein at the sites where we observe the most intense immunocytochemical labeling. Instead, our data may indicate that these receptors are synthesized in the somata and transported to the synaptic terminals. To date, although local translation of membrane proteins, heat-shock proteins, anti-oxidant proteins at distal axons and their terminals has been reported 27 28 , cytoskeletal proteins remain dominant among the axonally synthesized proteins 27 . The slow axoplasmic transport 29 and limited half-life of cytoskeletal proteins 30 31 render local synthesis a more effective means to deliver these proteins to distal axons or growth cones 32 33 . We do not know the half-life of serotonin receptor proteins in crayfish neurons, but the relatively stable lifespan (half-life >100 hours) of 5-HT_1A and 5-HT_2A receptors in rat 34 provides a reference. Secondly, therefore, the sparsity of receptor mRNA at sites where protein is most intensely labeled could be an outcome of a relatively short half-life of mRNA compared to the receptor protein. Indeed, in cultured P11 cells derived from rat pituitary tumors, 5-HT_2A mRNA has an average half-life of only 70 minutes 35 . One does not, therefore, necessarily expect to see co-localization of receptor mRNA and protein.

Our results and those of Spitzer et al. 24 demonstrate intense cytoplasmic labeling for the 5-HT_1α and 5-HT_2β receptors. Spitzer et al. have suggested that this labeling is likely to represent newly synthesized or recycled receptor. While this is quite possible, an alternative possibility is that at least some of these receptors may be functional as cytoplasmic proteins. One hypothesis is that serotonin may be taken into cells via the serotonin transporter, where it could bind to cytoplasmic receptors to initiate specific functions. It has been proposed that serotonin may act intracellularly as a growth or transcriptional regulator 36 37 38 39 , although the specific actions are speculative.

To explore the possibility of a cytoplasmic role for these serotonin receptors, we localized sites of SERT labeling in the crayfish brain. Interestingly, the pattern of labeling for the serotonin transporter is similar in many regions to the distribution of receptor labeling. For instance, the medial giant cells that contain intense cytoplasmic labeling for the 5-HT_2β receptor also label strongly for SERT. Likewise, cells in clusters 9, 10 and 11 contain intense labeling for both the 5-HT_1α receptor and SERT, while the LPZs are weakly labeled for both 5- HT_1α and SERT. These similarities between labeling for SERT and cytoplasmic 5-HT_1α and 5-HT_2β receptors may suggest a nontraditional role for serotonin in these neurons; serotonin may be taken into cells via the transporter, where it then activates cytoplasmic receptors. The large size of the medial giant cells makes these neurons particularly tractable for testing this hypothesis.

The distribution of 5-HT_1α and 5-HT_2β receptors in somata residing in several cell clusters, as well as in many synaptic neuropils (for example, protocerebral bridge, olfactory and accessory lobes, anterior and posterior median protocerebral neuropils) suggests that serotonin mediates a variety of functions in the crustacean brain. Previous studies have implicated 5-HT_1α receptors in the different degrees of social dominance in crayfish 40 and in diurnal rhythms in the eyestalk 41 , functions that are likely to involve higher order brain pathways. Our studies and those of Spitzer et al. 24 suggest a possible role for 5-HT_1α receptor in processing of olfactory information, as well as in higher order integrative functions mediated by the accessory lobes 8 9 . Less is known about possible physiological actions of 5-HT_2β receptor, although these have been localized to the processes and somata of lateral giant neurons, which are involved in the tail-flip escape response. Our present studies demonstrate the presence of 5-HT_2β receptor immunoreactivity in the cytoplasm of the medial giant cells, which also are part of the tail-flip circuitry 42 43 .

The primary goal of our studies was to examine the distribution of 5-HT_1α and 5-HT_2β receptors in the lineage of precursor cells that is responsible for the production of neurons in the adult crayfish brain, and to relate these findings to the effects of serotonin on the cell cycle of each generation of precursors. Exposure of intact animals to serotonin increases BrdU incorporation in the LPZ, which contains the late second- and third-generation neuronal precursors, but not into the first-generation precursors or their migrating daughters. This lineage will produce neurons that will differentiate into projection neurons that innervate the olfactory and accessory lobes. Pharmacological experiments with the 5-HT_1α receptor agonist QMS and the 5-HT_2β receptor antagonist MMS are consistent with the effects of serotonin on neuronal precursors, and with the distribution of these receptors in the neurogenic lineage.

These data are of additional interest because many features of adult neurogenesis are evolutionarily conserved. In mammalian and decapod crustacean brains, new neurons are added throughout life to the primary olfactory processing areas. The first generation neuronal precursors, which have glial properties, are located in specialized vascularized niches; their daughters migrate to sites where they will proliferate again and their progeny will differentiate into neurons 10 44 45 46 . Further, the timing and rate of proliferation in the adult crustacean brain, as in the mammalian brain, are influenced by circadian signals, age, diet, environmental enrichment, nitric oxide and serotonin 11 47 .

Serotonin is a particularly potent regulator of neurogenesis during both embryonic and adult life in mammals 48 49 50 51 and crustaceans 14 15 16 . Depletion of brain serotonin results in a decrease in production of new neurons in the dentate gyrus and the subventricular zone of adult rats 4 , and in clusters 9 and 10 containing local and projection interneurons in the olfactory pathway of lobsters 3 14 . The general influence of serotonin in the regulation of neurogenesis is therefore conserved across species.

The effects of serotonin on neurogenesis are mediated via its receptors. In mammals, several receptor subtypes are involved in regulating neuronal proliferation in the subgranular zone of the hippocampus and the subventricular zone that contributes new neurons to the olfactory bulb. By various accounts, the activation of 5-HT_1A and 5-HT_2C receptors increases neurogenesis in the subventricular zone/olfactory bulb, while 5-HT_1A, 5-HT_1B, 5-HT₂, 5-HT_2A, and 5-HT_2C receptors have been implicated in diverse effects on neuronal proliferation in the subgranular zone/hippocampus 52 53 54 55 56 57 58 . While it is apparent that there are differential actions of these receptors during distinct phases of neurogenesis in these brain regions, methodological disparities and contradictory results hinder clarity regarding these events. The interpretation of 5-HT receptor actions in mammalian systems also have been complicated by the complexity of the precursor cell lineage that produces adult-born neurons, the fact that multiple precursor generations coexist in neurogenic niches, and the relative scarcity of type 3 cells, which are a key population 55 . These challenges underscore the potential value of addressing the fundamental question of lineage-dependent influences of serotonin using the crayfish model, where precursor cell generations are spatially separated and quantitative changes are easily assessed. Furthermore, the use of diverse species to address important questions about adult neurogenesis is likely to result in a broader understanding of specific issues, and of how evolutionary processes may have shaped the vertebrate/mammalian condition.

In the crayfish brain, 5-HT_1α and 5-HT_2β receptors first appear in the late second-generation neuronal precursors at the end of their tangential migration across the ventral surface of the brain, as they reach the proliferation zones where olfactory interneurons will proliferate and differentiate. Neither the first generation precursor cells in the niche nor their daughters in the proximal and medial parts of the migratory stream express 5-HT_1α receptor. Further, the level of 5-HT_1α mRNA expression in the proliferation zones appears to be lower than in the mature cluster 10 neurons. This pattern suggests that after the first expression of 5-HT_1α receptor in the late second-generation precursors, receptor expression increases as their progeny differentiate into neurons. The functional assays using the 5-HT_1α agonist QMS corroborate the receptor localization, as QMS upregulates BrdU incorporation only in the proliferation zones of cluster 10, not among the niche precursors or in their daughters during migration to the proliferation zones. The story with the 5-HT_2β receptor appears very similar, in that 5-HT_2βCrust immunoreactivity is confined to the late second- and third-generation cells in the distal migratory streams and proliferation zones of cell clusters 9 and 10, as is the action of the 5-HT_2β antagonist MMS on cell proliferation in this system. This sequence of events is reminiscent of a study of the maturation of precursor cells migrating in the rostral migratory stream, which shows temporal and spatial regulation in the appearance of GABA and glutamate receptors, with GABA_A receptors expressed first in the tangentially migrating class 1 cells 59 , followed by AMPA receptors 60 ; during the next phase of maturation, the radially migrating class 2 cells express NMDA receptors.

Another interesting aspect of the 5-HT_1α receptor labeling pattern is the finding that only some of the cells in the distal migratory stream express the receptor. This variability may reflect heterogeneity in the spatiotemporal expression of the receptor in different cells, with the possible end point that all second-generation precursors will contain 5-HT_1α receptor. Alternatively, other classes of 5-HT receptors that were not assessed in this study may be involved in serotonin's action on cells in the proliferation zones. Finally, it may be that only a portion of the late second-generation precursors are sensitive to serotonin. It is known that there are at least two types of local interneurons in cluster 9 expressing either the neurotransmitter orkokinin or allatostatin-like peptide 10 . Similarly, there are two functional categories of projection neurons, those that innervate the olfactory lobe and others that innervate the accessory lobe 13 61 . The presence or absence of serotonin receptors in the late second-generation precursors may therefore reflect distinctive chemical/hormonal sensitivities involved in diverging cell fates.

Freshwater crayfish (carapace length 4 to 20 mm) P. clarkii (Malacostraca, Decapoda, Astacidae) of both sexes were obtained from Carolina Biological Supply Company (Burlington, NC, USA) and maintained at 21°C in aquaria with recirculating artificial pond water and a light:dark cycle of 12:12 hours. Because the basal rate of neurogenesis decreases with increasing age/size 25 , within each experiment the animals were carefully size-matched to within 1 to 2 mm carapace length.

Experimental animals were treated with serotonin creatinine sulfate (10^-9 M; Sigma, St Louis, MO, USA, no. H7752) and BrdU (2 mg/ml; Sigma, no. B5002) in pond water for 18 hours before sacrificing. Control animals were incubated in pond water with BrdU only (2 mg/ml) during the same time period. BrdU is incorporated into the DNA of replicating cells during the S phase of the cell cycle and can be visualized immunocytochemically 62 . The dose of serotonin used was based on previous results, as was the duration of application 15 . The exposure time was determined by the window required to observe changes in the proliferation of neuronal precursors.

To determine whether 5-HT_1α and 5-HT_2β receptors are involved in the serotonergic regulation of cell proliferation in the brains of P. clarkii, live animals were treated separately with QMS (a specific P. clarkii 5-HT_1α agonist; Sigma, Q1004), or MMS (a specific P. clarkii 5-HT_2β antagonist; Sigma, M149) for 2 hours, then with a mixture of QMS or MMS together with 2 mg/mL BrdU in pond water for another 8 hours before sacrificing. These pharmacological agents were selected based on previous studies of their superior specificity and efficacy on P. clarkii 5-HT_1α and 5-HT_2β receptors 20 . They were applied to intact animals in order to avoid artifacts associated with ex vivo systems. Control animals of the same size were treated with only BrdU for 8 hours.

Total RNA was extracted from crayfish tissues with TRIzol reagent according to the manufacturer's instructions (Invitrogen™ Life Technologies, USA). Reverse transcription was performed following the RNA extraction, and first-strand cDNA was synthesized with a random hexamer and SuperScript III reverse transcriptase (Invitrogen). PCR amplification of cDNA was performed using REDTaq ready mix PCR reaction mix (Sigma, no. R2523). The forward and reverse primers of 5-HT_1α, 5-HT_2β, and 18S ribosomal RNA (18S) were designed based on published sequences for P. clarkii [GenBank:EU131667, EU131666, and AF436001]. These were: 5'-AGAACACGACGAGCGATGA-3' and 5'-GCCAAGAATGACGGAAGTAA-3' (5-HT_1α); 5'-GATCTGTCCGCTGGAAGAAG-3' and 5'-ACCTGAAGCTCGAGTCGTGT-3' (5-HT_2β); and 5'-CTTCTTAGAGGGATTAGCGG-3' and 5'-TACGGAAACCTTGTTACGACTT-3' (18S).

Sense and anti-sense specific cRNA probes (riboprobes) for both 5-HT_1α and 5-HT_2β receptors were generated by first preparing modified PCR products similar to those described above where forward and reverse 5-HT receptor-specific primers also contained RNA polymerase initiation site sequences for T3 and T7 RNA polymerase, respectively. These modified PCR products were then used as templates, together with a DIG-labeling kit (Roche Molecular Biochemicals, Mannheim, Germany), to generate sense riboprobes with T3 RNA polymerase or anti-sense riboprobes with T7 RNA polymerase.

For whole mount in situ hybridization, dissected brains were fixed overnight at 4°C (4% w/v paraformaldehyde in RNAse-free 0.1 M phosphate buffer (PB; 20 mM NaH₂PO₄, 80 mM Na₂HPO₄; pH 7.4), and washed with PBTx (RNAse-free PB with 0.3% Triton X-100). In some cases endogenous peroxidase activity was quenched with 2% H₂O₂ treatment for 20 minutes. The brains were then acetylated in 0.1 M triethanolamine/acetic anhydride for 15 minutes, rinsed several times in PB and then in 2× SSC (SSC = 0.15 M NaCl, 0.015 M Na-citrate) for 5 minutes each. They were subsequently submerged for at least 16 hours at 70°C in hybridization buffer (50% formamide (Invitrogen), 5× SSC, 2% Blocking Reagent (Roche, no. 11096176001), 0.02% SDS, and 0.1% N-laurylsarcosine) with riboprobe concentrations of approximately 20 ng/ml. Sequential post-hybridization washes were performed at 70°C using hybridization buffer, 2× SSC, and finally 0.2× SSC. Whole mounts were then blocked in 0.1 M maleate buffer (pH 7.0) containing 1% Blocking Reagent and incubated in sheep anti-DIG-Fab fragments conjugated to alkaline phosphatase (0.15 U/ml; Roche, no. 11093274910) in maleate/blocking buffer for 1.5 hours. After rinses in maleate buffer and then in 0.1 M TRIS, 0.15 M NaCl, pH 9.5, brains were developed in the dark in substrate solution (pH 9.5) containing 0.1 M TRIS, 0.15 M NaCl, 50 mM MgCl₂, 2 mM levamisole, 0.4 mM 5-bromo-4 chloro-indolylphosphate (BCIP) and 0.4 mM nitroblue tetrazolium chloride (NBT) (both from Roche) to yield a dark purple color. Development time ranged from 4 to 13 hours until the desired color was achieved. Brains were rinsed in PB and images captured with a stereomicroscope (Nikon SMZ1500) or flat mounted in Clear-Mount (EMS, Hatfield, PA, USA) and then captured with a Nikon Eclipse 80i. Some post-hybridization brains were further immunolabeled by incubation with mouse anti-GS (1:100; BD Biosciences Pharmingen, San Jose, CA, USA, no. 610517) followed by goat anti-mouse IgG-HRP (1:600; Jackson Immunoresearch, West Grove, PA, USA, no. 115-035-166) or donkey anti-mouse IgG-Cy3 (1:100; Jackson Immunoresearch, no.711-165-152). HRP immunolabeling was visualized with the peroxidase substrate diaminobenzidine (DAB) (Sigma). In some cases post-hybridization brains were cryoprotected in 30% sucrose followed by cryosectioning into 20 μm sections.

For in situ hybridization in combination with BrdU immunofluoresence, fluorescence in situ hybridization was performed on whole mount brains. Following the blocking step in TRIS-NaCl buffer containing 0.5% blocking reagent (TNB; PerkinElmer, Boston, MA, USA) for 30 minutes at 21°C, the post-hybridization brains were incubated for 16 hours at 4°C in anti-DIG-POD (1:50 in TNB; Roche, no. 11207733910). Subsequently, brains were rinsed in TNT (0.1 M TRIS, 0.15 M NaCl, 0.05% Tween-20), and incubated for 1 hour in Cy3-tyramide (1:50; PerkinElmer, SAT704A). Brains were then incubated sequentially with mouse anti-BrdU antibody (1:50 in TNB; BD Biosciences, no. 34758) and then with Cy5-conjugated donkey anti-mouse IgG secondary antibody (1:100 in TNB; Jackson Immunoresearch, no. 715-175-151). The brains were rinsed in TRIS-NaCl and coverslips were applied with Fluoromount (EMS). Control brains were hybridized with sense riboprobes.

For immunocytochemistry, brains were dissected in cold crayfish saline (205 mM NaCl, 5.4 mM KCl, 34.4 mM CaCl₂, 1.2 mM MgCl₂ and 2.4 mM NaHCO₃) and then fixed (4% w/v paraformaldehyde in PB) for 16 hours at 4°C. Immunofluorescence methods for 5-HT_1α and 5-HT_2β receptors were modified from previously published protocols 19 23 . In brief, fixed brains were dehydrated through an ethanol series (10 minutes each: 30, 50, 70, 80, 90, 95, 100%), then rehydrated (100, 70, 50, 30% ethanol). Preparations were washed in PBTx and then incubated 16 hours at 4°C with 2 μg/ml rabbit anti-5-HT_1α (5-HT_1crust 19 24 ) or with 0.5 to 1 μg/ml rabbit anti-5-HT_2β (5- HT_2βCrust 19 ) antibody (gift from D Baro, Department of Biology, Georgia State University). After a 2-hour wash in PBTx, the brains were incubated in donkey anti-rabbit IgG-Cy2 (1:100 in PBTx; Jackson Immunoresearch, no. 711-225-152) for 16 hours at 4°C. The preparations were then washed in PBTx, dehydrated sequentially in a graded ethanol series up to 100%, cleared in methyl salicylate (Sigma, M-6752) and mounted in DPX (Fluka Chemie, Buchs, Switzerland, no. 44581). The specificity of the primary antibody for 5-HT_1crust was confirmed by incubating brains with antiserum that was preadsorbed with the peptide KDPDFLVRVNEHKKCLVSQD (gift from D Baro, Department of Biology, Georgia State University) which was used to raise the antibody. The primary antibody for 5- HT_2βCrust was raised against peptide DRFLSLRYPMKFGRHKTRRR and its specificity was confirmed by Clark et al. 19 with antiserum that was preadsorbed with this peptide.

For double labeling of 5-HT_1α or 5-HT_2β receptors and BrdU, the fixed brains from BrdU-treated animals were first treated with 2N HCl for 45 minutes, and then a mixture of 5-HT_1crust or 5- HT_2βCrust and mouse anti-BrdU (1:50; BD Biosciences) was applied. The secondary antibodies were donkey anti-rabbit IgG-Cy2 and donkey anti-mouse IgG-Cy3 (1:100). For double labeling of 5-HT_1α or 5- HT_2β receptors and GS, the brains were sequentially incubated first with a mixture of rabbit 5-HT_1crust or 5- HT_2βCrust and mouse anti-GS, and then a mixture of donkey anti-rabbit IgG-Cy2 and donkey anti-mouse IgG-Cy5. The GS antibody served as a glial marker, as reported in previous studies in crustaceans 10 63 64 .

For labeling of SERT, an antigen retrieval step of 2N HCl for 45 minutes was added before incubation in mouse monoclonal anti-SERT antibody (1:1000; Advanced Targeting Systems, San Diego, CA, USA, no. AB-N09). The retrieval step significantly increases the signal to noise ratio. The immunogen was a peptide from the fourth extracellular domain of rat SERT. The specificity of anti-SERT was confirmed by preadsorption with peptide (CEMRNEDVSEVAKDA; Advanced Targeting Systems, no. PR-03). For double-labeling with BrdU, rat anti-BrdU (Accurate Chemical Co., Westbury, NY, USA, no. OBT0030G) with donkey-anti-rat IgG-Cy3 (Jackson Immunoresearch, no. 712-165-153) was used.

Before mounting brains, color images of the in situ hybridizations were captured with a stereomicroscope (Nikon SMZ1500). After mounting, the preparations were photographed with a Nikon Eclipse 80i microscope to obtain highly magnified images. Fluorophore-labeled specimens were visualized with a Leica TCS SP laser scanning confocal microscope equipped with argon 488 nm, krypton 561 nm and helium-neon 633 nm lasers. Serial optical sections were taken at intervals of 1 μm and saved as both three-dimensional stacks and two-dimensional projections. Image preparation, assembly and analysis were performed in Photoshop 7 (Adobe Systems, San Jose, CA, USA). Only the color balance and contrast of the images were adjusted.

The numbers of BrdU-labeled cells in the niche, migratory streams and the LPZs in cluster 10 were blind counted by individual observers as previously described 25 . In brief, a single optical section was projected onto the monitor and the labeled cells traced onto a transparent sheet. This was repeated for each optical section and the cell profiles then counted from the sheets. All data are presented as mean ± standard error of the mean. Comparisons between different groups of animals were made with Student t-tests, or one-way ANOVA analysis followed by Tukey's multiple comparison tests, using Prism software (GraphPad Software, San Diego, CA, USA). Labeled cells in cluster 9 were not assessed quantitatively in the present studies because the MPZ of cluster 9 is less well-defined than the LPZ in cluster 10. The proliferation zone that contributes cells to cluster 9 (MPZ) is located slightly posteromedially from cluster 9 and merges with migratory cells, while the proliferation zone contributing cells to cluster 10 (LPZ) is highly localized within cluster 10 (see Figure 1), providing a more reliable quantitative assay.