Melatonin Synthesis: Acetylserotonin O-Methyltransferase (ASMT) Is Strongly Expressed in a Subpopulation of Pinealocytes in the Male Rat Pineal Gland

Abstract

The rat pineal gland has been extensively used in studies of melatonin synthesis. However, the cellular localization of melatonin synthesis in this species has not been investigated. Here we focus on the localization of melatonin synthesis using immunohistochemical methods to detect the last enzyme in melatonin synthesis, acetylserotonin O-methyltransferase (ASMT), and in situ hybridization techniques to study transcripts encoding ASMT and two other enzymes in melatonin synthesis, tryptophan hydroxylase (TPH)-1 and aralkylamine N-acetyltransferase. In sections of the rat pineal gland, marked cell-to-cell differences were found in ASMT immunostaining intensity and in the abundance of Tph1, Aanat, and Asmt transcripts. ASMT immunoreactivity was localized to the cytoplasm in pinealocytes in the parenchyma of the superficial pineal gland, and immunopositive pinealocytes were also detected in the pineal stalk and in the deep pineal gland. ASMT was found to inconsistently colocalize with S-antigen, a widely used pinealocyte marker; this colocalization was seen in cells throughout the pineal complex and also in displaced pinealocyte-like cells of the medial habenular nucleus. Inconsistent colocalization between ASMT and TPH protein was also detected in the pineal gland. ASMT protein was not detected in extraepithalamic parts of the central nervous system or in peripheral tissues. The findings in this report are of special interest because they provide reason to suspect that melatonin synthesis varies significantly among individual pinealocytes.

The pineal gland is a neuroendocrine organ that transforms photoperiodic information into a hormonal melatonin signal (1, 2). The term pineal gland, as used in the rodent literature, commonly refers to one element of the pineal complex, the superficial pineal gland; in addition, the pineal complex includes the deep pineal gland and a stalk connecting the two (3, 4). In the rat, all three parts of the pineal complex have been shown to express genes required for melatonin synthesis (5–9). Melatonin is synthesized from serotonin in a two-step process by the sequential action of the pivotal enzymes aralkylamine N-acetyltransferase (AANAT) and acetylserotonin O-methyltransferase (ASMT) (1, 10) (Figure 1). Aanat and Asmt are expressed at very high levels only in the pineal gland and are essential to the well-established physiological function of the pineal gland, namely melatonin synthesis. Pineal synthesis of melatonin accounts for circulating levels of melatonin, which increase at night due to increase in AANAT activity (10), whereas ASMT activity varies little on a 24-hour basis in the mammalian pineal gland.

Figure 1.

Melatonin synthesis. Melatonin is synthesized from tryptophan by the sequential action of TPH1, dopa decarboxylase (DDC; also known as aromatic L-amino acid decarboxylase), AANAT, and ASMT.

By definition, the pinealocyte is the principal melatonin-producing cell type of the pineal gland. Pinealocytes comprise more than 95% of the cells in the rat pineal gland based on morphological characteristics (11), which has led to the assumption by workers in the field that 95% of pineal parenchymal cells are functionally identical pinealocytes that are biochemically similar and regulated by the same mechanisms. However, based on ultrastructural characteristics, eg, the density of the cytoplasm, pinealocytes have been divided into two subtypes, both of which are present in some mammals (4). Although this does not apply to the rat, some degree of heterogeneity is known to exist within the rat pinealocyte population, as evidenced by marked cell-to-cell differences in S-antigen (SAG) immunoreactivity in the intact gland (12) and in cultured pinealocytes (13), the latter of which also exhibit marked heterogeneity in membrane currents and calcium responsiveness (14, 15).

Although it is clear that rat pinealocytes are heterogeneous, nothing is known about the cellular site of melatonin synthesis. Specifically, immunocytochemical investigations on the presence of melatonin synthesizing enzymes, which have been done for the human, sheep, and bovine pineal gland using species-specific antisera (16, 17), have not been performed on the rat pineal gland. The rat is the most widely used experimental species in pineal research, and because of this, it is of fundamental importance to define the cellular localization of melatonin synthesis. Here we characterize the first antirat ASMT antisera and have used them in conjunction with nonisotopic in situ hybridization methods to identify cells expressing genes linked to melatonin synthesis. Our findings are consistent with the view that pineal melatonin synthesis varies considerably between major subpopulations of rat pinealocytes and provide reason to consider ASMT as a novel cellular marker.

Materials and Methods

Animals

Adult male Sprague Dawley rats (150–200 g) were obtained from Charles River with the exception of those used for semiquantitative Western blot analyses, which were obtained from Taconic Farms. Animals were housed under a 12L:12D light-darkness (LD) schedule with food and water ad libitum. For immunohistochemistry, animals were perfusion fixed in 4% paraformaldehyde, the brain and eyeballs were removed and immersed in the same fixative for 24 hours, subsequently cryoprotected in 25% sucrose, and frozen in crushed solid CO₂; alternatively, the fixed tissue was embedded in paraffin. For in situ hybridization, animals were decapitated, and brains were removed and frozen on crushed solid CO₂. For Western blot analysis, animals were decapitated, brains and peripheral tissues were dissected, and samples were frozen in crushed solid CO₂. For each experiment, the Zeitgeber time (ZT) of euthanasia is given in the figure legend. During the dark period, animals were killed under dim red light. All experiments with animals were performed in accordance with the guidelines of European Union Directive 86/609/EEC (approved by the Danish Council for Animal Experiments) and the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Cloning and sequencing of rat ASMT

Full-length rat Asmt mRNA (1520 bp) was cloned as follows. Amplification primers were designed using the reference genomic sequence. The forward primer was 5′-GAGGCTAGGATAGCAAGATGG-3′; the reverse primer was: 5′-ACTGTGACATCACTTCCTGC-3′. Herculase II Fusion polymerase (Agilent) was used to amplify the rat pineal gland cDNA to generate amplicons, which were then cloned into the pCR-Blunt vector (Thermo Fisher) and sequenced completely in both directions (Macrogen). A 14-residue peptide sequence was extracted from the portion of rat ASMT that was downstream from the aberrant stop codon generated by the original NM_144759.1. This peptide, EGWERQASDYRNLA corresponding to position 308–321 on rat ASMT protein (NP_653360.2), was used to generate two antisera against rat ASMT.

Production of antisera

Antisera were custom produced by GenScript. Two New Zealand rabbits (numbered 8900 and 8901) were injected with a synthetic rat ASMT_308–321 peptide conjugated to keyhole limpet hemocyanin (Table 1). Preimmune bleedings and postimmunization bleedings were obtained from each rabbit.

Table 1.

Antibody Information

Protein Target	Antigen Sequence	Name of Antibody	Manufacturer or Person Providing the Antibody	Species Raised (Monoclonal or Polyclonal)	Dilutions Used
ASMT	EGWERQA SDYRNLA	Antirat ASMT308–321 8900	Dr David C. Klein	Rabbit polyclonal	1:200 1:500 1:1000
		Antirat ASMT308–321 8901
SAG	Bovine retinal S-antigen	MAbA9-C6	Dr Larry A. Donoso	Mouse monoclonal	1:200
Actin	SGPSIVHRKCF	Antiactin, clone AC40	Sigma-Aldrich, catalog number A4700	Mouse monoclonal	1:1000
TPH	FANOILSYAELDADHPG FKDPYR	PH8	Millipore, catalog number MAB5278	Mouse monoclonal	1:2000 1:5000

For purification purposes (18), diluted samples of antisera (1:500 in PBS with 1% bovine serum albumin and 0.3% Triton X-100) were incubated (2 d, 4°C) with free-floating nonpineal rat brain sections; a volume of 20 mL diluted serum was incubated with sections of one whole perfusion fixed brain (600 sections, 40 μm thick) (Supplemental Figure 1). Alternatively, diluted samples of antisera (1:500 in PBS with 1% BSA and 0.3% Triton X-100) were incubated (2 d, 4°C) with a curtain Western blot containing electrophoresed protein samples from extraepithalamic brain areas; a volume of 5 mL diluted serum was incubated with a 50-cm² membrane containing a total of 1 mg blotted brain protein sampled from the neocortex, striatum, hypothalamus, cerebellum, and brainstem (Supplemental Figure 1).

Immunohistochemistry

Cryostat sections used for immunohistochemistry were either mounted on slides (section thickness 14 μm) or free floating (section thickness 40 μm); alternatively, paraffin sections (5 μm thick) were used. Sections were washed in PBS for 3 × 5 minutes and blocked in 5% normal swine serum diluted in PBS for 30 minutes. This was followed by incubation overnight in the primary rabbit antirat ASMT antisera diluted 1:200, 1:500, or 1:1000 (Table 1) in PBS with 1% bovine serum albumin and 0.3% Triton X-100 at 4°C. For specificity controls (Supplemental Figure 2), adjacent sections from the same animals were incubated in antiserum, preimmune serum, or preabsorbed antiserum and reacted in parallel. For preabsorption, the diluted antisera were incubated with the immunogenic peptide (ASMT_308–321, 100 μg/mL) for 2 days at 4°C. Sections were subsequently washed 3 × 10 minutes in PBS with 0.25% bovine serum albumin and 0.1% Triton X-100 followed by incubation for 1 hour in secondary biotinylated swine antirabbit IgG (Dako) diluted 1:500 in the same buffer. The sections were washed 3 × 5 minutes in PBS with 0.1% Triton X-100 and incubated for 45 minutes in ABC Vectastain (Vector Laboratories) diluted 1:100 in the same buffer. The chromogenic reaction was performed in 1.4 mM diaminobenzidine (Sigma) and 0.01% H₂O₂ in 0.05 M Tris (pH 7.6) for 10 minutes. Sections were washed in deionized water, dried, and coverslipped in Pertex (Histolab). For the detection of tryptophan hydroxylase (TPH), the same protocol was followed with the following modifications: sections were incubated in a primary mouse monoclonal anti-TPH antibody (PH8; Millipore) (19, 20) diluted 1:5000 (Table 1) and in secondary biotinylated rabbit antimouse IgG (Dako) diluted 1:500.

For fluorescent detection of the immunoreaction, sections were incubated in the primary rabbit antirat ASMT antisera diluted 1:200 (Table 1) followed by incubation in an Alexa Flour 488-conjugated donkey antirabbit or Alexa Flour 568-conjugated goat antirabbit secondary antisera (Invitrogen) diluted 1:300. For colocalization studies with SAG, sections were incubated in a combination of rabbit antirat ASMT antisera diluted 1:200 and mouse monoclonal anti-S-antigen (MAbA9-C6) (21) diluted 1:200 (Table 1). For colocalization studies of TPH1 and ASMT, sections were incubated in a combination of rabbit antirat ASMT antisera diluted 1:500 and a mouse monoclonal anti-TPH antibody (PH8; Millipore) (19, 20) diluted 1:2000 (Table 1). Sections were subsequently incubated in Alexa Flour 568-conjugated goat antirabbit and Alexa Flour 488-conjugated donkey antimouse secondary antisera (Invitrogen) both diluted 1:300. Finally, sections were stained in 4′,6′-diamidino-2-phenylindole and coverslipped in fluorescent mounting medium (Dako).

Western blotting

Frozen tissue samples were homogenized in 2× Laemmli at room temperature (22), boiled for 10 minutes, and centrifuged for 1 hour at 13 000 × g at 4°C. The supernatant was collected and protein concentration was determined by use of the bicinchoninic acid protein assay (Pierce), and samples were subsequently diluted and boiled in 2× Laemmli with 5% 2-mercaptoethanol (final sample content) for 5 minutes. Protein (50 μg/lane) was run in a NuPAGE Bis-Tris 4%–12% gel and transferred to a nitrocellulose membrane by use of the Sure-Lock minicell system (Invitrogen). The membrane was blocked in 2% skim milk in PBS buffer for 30 minutes. For detection of ASMT, the membrane was incubated in primary rabbit antirat ASMT antisera diluted 1:500 in 2% skim milk overnight at 4°C (Table 1). For specificity controls (Supplemental Figure 3), the same samples were run side by side, and parallel strips of the same membrane were incubated in ASMT antiserum, preimmune serum, or absorbed antiserum and reacted in parallel. For preabsorption, the diluted antisera were incubated with the immunogenic peptide (ASMT_308–321, 50 μg/mL) for 2 days at 4°C. The membrane was washed in PBS for 3 × 10 minutes and incubated for 1 hour in secondary biotinylated swine antirabbit IgG (Dako) diluted 1:500 in 2% skim milk. The membrane was subsequently washed in PBS 3 × 5 minutes and incubated for 45 minutes in ABC Vectastain (Vector Laboratories) diluted 1:100. The blot was developed in 1.4 mM diaminobenzidine (Sigma) and 0.01% H₂O₂ in 0.05 M Tris (pH 7.6) followed by washing in deionized water.

For the detection of TPH, the same protocol was followed with the following modifications: the membrane was incubated in a primary mouse monoclonal anti-TPH antibody diluted 1:5000 (Table 1) and in secondary biotinylated rabbit antimouse IgG (Dako) diluted 1:500. Protein size was estimated by comparison with Precision Plus kaleidoscope prestained protein standards (Bio-Rad Laboratories) or HiMark prestained protein standards (Life Technologies). Semiquantitative Western blot analyses were performed using an antibody against actin (Sigma) in a dilution of 1:1000 (Table 1) and the Odyssey infrared fluorescent detection system (LI-COR Biosciences) as previously described (23).

In situ hybridization

Detection of Asmt and Tph1 transcripts

Riboprobe templates were generated from nighttime pineal cDNA (24) by the use of PCR. Primers were designed to amplify positions 539–1116 on rat Asmt mRNA (NM_144759.2) and 564–1449 on rat Tph1 mRNA (NM_001100634.2), respectively. The Tph1 riboprobe template was based on a previously published sequence (25). PCR products were purified by electrophoresis followed by gel extraction (QIAGEN), subsequently cloned into pGEM-T Easy vectors (Promega) and amplified in DH5-α cells (Invitrogen). Direct sequencing of plasmid inserts confirmed sequence identity (Macrogen). Plasmids were subsequently used as templates for PCRs with insert-specific primers extended with a 5′ T3 (forward primer) or a 3′ T7 (reverse primer) viral promoter sequence. For Asmt, the following primer sequences were used: 5′-CGCGCAATTAACCCTCACTAAAGGGACTTTGACCTCTCACGCTTCC-3′/5′-GCGCGTAATACGACTCACTATAGGGCAAAACTTCCTGTCCCCGTA-3′; for Tph1, the following primer sequences were used: 5′-CGCGCAATTAACCCTCACTAAAGGGATCGCAGAGCTGGCTACTACA-3′/5′-GCGCGTAATACGACTCACTATAGGGACTGGGCCACCTGCTGACTCTA-3′.

PCR was performed on 150 ng plasmid in a total volume of 50 μL by use of Taq polymerase (Roche) in accordance with the manufacturer's instructions with the following temperature protocol: 94°C for 5 minutes, five cycles of 94°C for 30 seconds, 60°C for 45 seconds, and 72°C for 60 seconds, followed by 25 two-step cycles of 94°C for 30 seconds and 72°C for 60 seconds, and finally 94°C for 10 minutes. PCR products were purified by agarose gel electrophoresis followed by gel extraction (QIAGEN). Transcription of riboprobes was performed using 100 ng purified PCR product as a template in a total volume of 10 μL by use of T7 RNA polymerase and digoxigenin (DIG) RNA labeling mix (Roche) at 37°C for 2 hours in accordance with the manufacturer's instructions. The DIG-labeled RNA probe was purified by use of RNA probe purification kit (Omega Biotek). The in situ hybridization procedure was modified from elsewhere (26). Cryostat sections (12 μm, mounted on slides) were thawed for 5 minutes and fixed for 10 minutes in 4% formaldehyde in PBS. Sections were washed for 2 × 5 minutes in PBS and acetylated in 0.25% acetic anhydride for 10 minutes. Sections were dehydrated in a graded series of ethanol (70%, 80%, 95%, 100%, and 95%, 1 minute each) and dried. The labeled probe (20 ng/μL) was mixed with RNA mix (5 μg/μL) (Sigma), heated at 65°C for 5 minutes, and chilled. Finally, the denatured probe was diluted in hybridization buffer (27) with sodium thiosulfate and sodium dodecyl sulfate (both 0.05%) to a final concentration of 500 ng (Tph1) or 1 μg (Asmt) RNA probe/mL.

Hybridization was performed in a humid chamber (50% formamide/4× saline sodium citrate [SSC]) at 50°C overnight. The sections were washed in 4× SSC for 4 × 5 minutes at room temperature, 2× SSC for 2 × 5 minutes at room temperature, 1× SSC for 5 minutes at room temperature, 0.5× SSC for 5 minutes at room temperature, and 0.1× SSC for 30 minutes at 65°C. Slides were subsequently cooled to room temperature in fresh 0.1× SSC and washed in Tris-buffered saline (TBS; 150 mM sodium chloride; 100 mM Tris, pH 7.5) for 2× 5 minutes. The sections were blocked in 5% normal goat serum diluted in TBS for 30 minutes and incubated in anti-DIG-AP (Roche) diluted 1:2000 in the same buffer at 37°C for 5 hours. Sections were washed in TBS for 3 × 3 minutes and in DIG buffer (100 mM Tris; 100 mM sodium chloride; 50 mM magnesium chloride, pH 9.5) for 2 × 3 minutes. Sections were subsequently incubated in 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate, toluidine salt stock solution (Roche) diluted 1:50 in DIG buffer with Levamisole (Dako) added in accordance with the manufacturer's instructions. After incubation for 4–12 hours in darkness, sections were washed in deionized water for 3 × 5 minutes and coverslipped in fluorescent mounting medium (Dako). For specificity control (Supplemental Figure 4), parallel tissue sections were hybridized in parallel with sense probes generated from the same template by use of T3 RNA polymerase (Roche).

Detection of the Aanat transcripts

Sagittal 12-μm cryostat sections were dehydrated, hybridized, and washed as previously described (24) with the exception that hybridization was performed with a 3′ biotinylated oligo probe (DNA Technology) at a concentration of 50 nM. The probe sequence, 5′-GTCCCAAAGTGAACCGATGATGAAGGCCACAAGACACC-3′, corresponding to position 409–372 on rat Aanat mRNA (NM_012818.2), was identical with that of a published probe, the specificity of which has been previously confirmed (8, 28, 29). After hybridization and washing in SSC, sections were rinsed in PBS with 0.1% Triton X-100 and incubated in ABC Vectastain (Vector Laboratories) diluted 1:100 in the same buffer for 45 minutes. Sections were washed in PBS with 0.1% Triton X-100 for 3 × 5 minutes and incubated in biotinylated tyramide containing 45 μM tyramine (Sigma) and 45 μM NHS-LC-biotin (Pierce) diluted in PBS with 0.005% H₂O₂ for 6 minutes. Sections were washed in PBS with 0.1% Triton X-100 for 2 × 5 minutes and 0.05 M Tris (pH 7.6) for 5 minutes followed by incubation in Streptavidin/AP (Dako) diluted 1:100 in 0.05 M Tris (pH 7.6) for 20 minutes. Sections were washed in 0.05 M Tris (pH 7.6) for 3 × 5 minutes followed by chromogenic development in Fast Red solution (Dako) according to the supplier's instructions for 30 minutes. Sections were washed in deionized water and coverslipped in fluorescent mounting medium (Dako).

Statistical analysis

Semiquantitative Western blot data were analyzed by use of a one-way ANOVA in GraphPad Prism 5.0 (GraphPad Software). A value of P < .05 was considered to represent statistical significance.

Results

Cloning and sequencing of rat Asmt

The rat Asmt probe and antirat ASMT antisera used in this study were based on a revised Asmt sequence we obtained by cloning and sequencing full-length rat Asmt. This sequence (EU741665.1) formed the basis of Asmt reference sequencing (NM_144759.2); the prior rat Asmt reference sequencing (NM_144759.1) did not fully correspond to the genomic Asmt sequence and encoded a truncated protein. We obtained multiple clones of a transcript that contained eight exons and matched the reference genomic sequence (Supplemental Figure 5). This transcript encodes a 432-amino acid peptide that is 53% and 71% identical with that of human and mouse ASMT, respectively. Rat ASMT contains 87 residues at its C terminus that are not present in the human and mouse proteins.

In addition to this eight-exon transcript, multiple clones were isolated that were missing exon 2 (160 bp), as confirmed by PCR and by RNA sequencing (30). In most cases, the absence of exon 2 resulted in a frame shift and severely truncated protein (variant 3; GenBank accession number KT820177) (Supplemental Figure 6). In one case, however, in addition to the absence of exon 2, there was a 13-bp extension of exon 1 that restored the open reading frame; this form was supported by RNA sequencing (30) and has been deposited in GenBank (variant 2; accession number KT820176) (Supplemental Figure 6).

Immunohistochemical analysis of ASMT protein in the rat pineal complex

The cellular localization of ASMT in the rat pineal gland was examined immunohistochemically by use of two novel rabbit antirat ASMT sera and sections of the rat brain containing the pineal complex (Figures 2 and 3), as detailed in Materials and Methods.

Figure 2.

Immunohistochemistry for detection of ASMT in sections of the superficial pineal gland of the rat. A, Sagittal cryostat section of the superficial pineal gland (low magnification) reacted by use of an enzymatic method for detection of ASMT. Immunopositive pinealocytes are present in all parts of the pineal parenchyma but not in the perivascular spaces, here indicated by an asterisk. Scale bar, 100 μm. B, Coronal paraffin section of the superficial pineal gland (high magnification) reacted by use of an enzymatic method for detection of ASMT. The reaction product is located in the cytoplasm. The staining intensity varies between individual pinealocytes. Arrow indicates a strongly stained pinealocyte, whereas the arrowhead indicates a weakly stained pinealocyte. The asterisk indicates a perivascular space. Scale bar, 20 μm. C, Coronal section of the superficial pineal gland reacted for fluorescent detection of ASMT. Immunopositive pinealocytes are present in all parts of the pineal parenchyma. Scale bar, 200 μm. Antisera generated in rabbit number 8900 (A and B) and number 8901 (C) were used for the reactions displayed; both antisera produced similar results in dilutions of 1:200 (enzymatic and fluorescent detection) and 1:1000 (enzymatic detection). For specificity controls and direct comparison of antisera, see Supplemental Figure 1, Supplemental Figure 2, and Supplemental Figure 8. Animals were killed at ZT18 (A and B) or during daytime (C).

Figure 3.

Immunohistochemistry for detection of ASMT in sections of the pineal stalk and deep pineal gland of the rat. A, Sagittal section of the pineal stalk. Densely stained pinealocytes are present in the pineal stalk (PS). Scale bar, 100 μm. B, Sagittal section of the deep pineal gland. Densely stained pinealocytes are seen in the deep pineal gland (DPG) between the habenular commissure (HC) and the posterior commissure (PC). Scale bar, 100 μm. DPG, deep pineal gland; HC, habenular commissure; PC, posterior commissure; PS, pineal stalk. C, Sagittal section of the pineal stalk in high magnification. Scale bar, 20 μm. D, Sagittal section of the deep pineal gland in high magnification. Scale bar, 20 μm. Arrowheads indicate pinealocytes with clearly stained cytoplasm. Antisera from rabbit numbers 8900 and 8901 produced similar results in dilutions of 1:200 and 1:1000. Antiserum generated in rabbit number 8900 was used for the reactions displayed. Animals were killed at ZT18.

The superficial rat pineal gland contained immunoreactive cells in the pineal parenchyma, whereas cells and cellular structures of the interstitial spaces containing interstitial cells, blood vessels, and nerve fibers were negative, strongly indicating specific staining of pinealocytes (Figure 2). Two approximately similar-sized populations of pinealocytes could be identified, one of which was densely labeled and the other negative (Figure 2B and Supplemental Figure 7). This markedly heterogeneous staining pattern was observed with two detection methods, enzymatic (Figure 2, A and B) and fluorescent (Figure 2C), and with two different antisera (Supplemental Figures 1 and 2). The ASMT immunoreactivity was located in the cytoplasm (Figure 2B); in some pinealocytes, the reaction product was predominantly located in a perinuclear position, whereas in others the protein was also seen to extend into pinealocyte processes. Pinealocytes exhibiting the strongest ASMT signal were randomly distributed in the pineal parenchymal cords and were not located close to or in a gradient around the perivascular spaces, suggesting that expression is not determined by structural features, such as nerve fibers or blood vessels. The specificity of the antisera was confirmed by incubation in preimmunization bleedings and preabsorption with the immunogenic peptide (Supplemental Figure 2).

Examination of other parts of the pineal complex revealed strongly stained ASMT-immunopositive pinealocytes throughout the pineal stalk extending from the rostral part of the superficial pineal gland; these strongly immunopositive pinealocytes could be followed ventrorostrally along the stalk to the deep pineal gland (Figure 3, A and B). In these structures, the heterogeneous staining pattern was also seen; furthermore, the reaction product was detectable in the cytoplasm (Figure 3, C and D).

Immunohistochemical colocalization studies of ASMT and SAG

SAG is regarded as a pinealocyte marker (12, 21, 31). Of special interest regarding the heterogeneous nature of the ASMT-immunopositive cells observed here is the similar pattern of SAG-immunopositive cells in the rat pineal gland (12). Accordingly, we performed colocalization studies, which revealed a subset of pinealocytes in the pineal complex that was positively stained for both SAG and ASMT, whereas others were immunopositive for only one of the proteins (Figure 4, A–D). The caudal part of the medial habenular nucleus also contained pinealocyte-like cells that were immunopositive for both proteins (Figure 4, E and F).

Figure 4.

ASMT colocalization with SAG in the rat pineal complex. A and B, Sagittal section of the superficial pineal gland reacted for the detection of ASMT (A) and SAG (B). C and D, Sagittal section of the pineal stalk reacted for detection of ASMT (C) and SAG (D). E and F, Sagittal section of the caudal part of the medial habenular nucleus reacted for detection of ASMT (E) and SAG (F). Scale bars, 20 μm. Colocalization is detectable in some cells (arrows), whereas other cells are positive for only one of the proteins detected (arrowheads). ASMT antiserum generated in rabbit number 8900 was used for the reactions displayed. Rats were killed at ZT6.

Tissue distribution of ASMT protein

The tissue distribution of the ASMT protein was analyzed by Western blotting (Figure 5 and Supplemental Figure 8A) and immunohistochemistry (Supplemental Figure 8B). A band of 45 kDa was detectable only in extracts of the pineal gland but not in other parts of the central nervous system or in any peripheral tissues analyzed; this corresponds to the predicted molecular mass of 45.6 kDa of rat ASMT (NP_653360.2). This analysis did not detect a protein corresponding to the predicted molecular mass (40.3 kDa) of the truncated isoform of ASMT (variant 2; GenBank accession number KT820176) in any tissue. Notably, when analyzing sagittal and coronal sections of the whole brain, ASMT was detected only in the epithalamus (Supplemental Figure 8B). It was not possible to specifically detect ASMT in the retina by immunohistochemistry or Western blot (Figure 5 and Supplemental Figure 3).

Figure 5.

Western blot for detection of ASMT in rat tissues. Animals were killed at ZT6 (D) or ZT18 (N). Two gels were run in parallel: one for the central nervous system (left) and one for peripheral tissues (right). Arrows indicate molecular weight markers. Rat ASMT (NP_653360.2) has a predicted molecular mass of 45.6 kDa; a 45-kDa is present only in the pineal gland. Antiserum generated in rabbit number 8901 was used for the blot displayed. For specificity controls and direct comparison of antisera, see Supplemental Figure 1, Supplemental Figure 3, and Supplemental Figure 8.

The pineal-restricted distribution of ASMT was confirmed using two different antisera, both of which detected a pineal-specific, 45-kDa band in Western blotting analyses (Supplemental Figure 8A); specificity was confirmed by incubation in preimmune sera and preabsorption with the immunogenic peptide (Supplemental Figure 3). Furthermore, purification of the antisera by exposure to epithalamic-free brain sections or a curtain blot of epithalamic-free brain protein enhanced Western blot results without altering the distribution of ASMT-immunoreactivity in the pineal gland (Supplemental Figure 1), thus confirming that the protein detected in immunohistochemical reactions corresponds to the 45-kDa ASMT band.

Diurnal analysis of pineal ASMT protein

Pineal ASMT was detected by Western blotting in both daytime and nighttime samples (Figure 6). Semiquantitative Western blot analyses on samples obtained throughout the 24-hour daily cycle (see Figure 8A) did not reveal significant daily changes in pineal ASMT protein levels (P = .08, one way ANOVA). Immunohistochemical analyses also showed the presence of ASMT protein in the pineal gland of animals sacrificed during daytime (ZT6) and nighttime (ZT18), respectively, and daily changes in staining pattern were not observed (Figure 6, B and C).

Figure 6.

Diurnal analysis of ASMT in the rat pineal gland. A, Semiquantitative Western blot analysis of ASMT in the rat pineal gland. Glands were obtained from rats housed under a 12-hour light (white part of panel), 12-hour darkness (black part of panel). The signal intensity of the 45-kDa ASMT-band was normalized against actin. The graph represents data from three to five independent experiments. The mean value for each time point with SEM is displayed. AU, arbitrary units. B and C, Immunohistochemistry for detection of ASMT in the superficial pineal gland of rats killed during day (B, ZT6) and night (C, ZT18). Immunoreactions were run in parallel. Scale bar, 20 μm. Antiserum generated in rabbit number 8900 was used for the displayed analyses.

Figure 8.

ASMT colocalization with TPH1 in the rat pineal complex. A and B, Sagittal section of the superficial pineal gland reacted for detection of ASMT (A) and TPH1 (B). C and D, Sagittal section of the deep pineal gland reacted for detection of ASMT (C) and TPH1 (D). Colocalization is detectable in some cells (arrows), whereas other cells are positive for only one of the proteins detected (arrowheads). ASMT antiserum generated in rabbit number 8900 was used for the reactions displayed. Rats were killed at ZT18. The antibody used for detection of TPH1 is further characterized in Supplemental Figure 9. Scale bars, 20 μm.

Cellular detection of transcripts encoding melatonin-producing enzymes in the pineal gland

Transcripts encoding the enzymes involved in melatonin synthesis have been detected in the pineal gland by use of in situ hybridization (3, 5, 6, 8, 25, 29, 32). These prior studies have used radiochemical in situ hybridization, the resolution of which is not sufficient to determine the cellular location of pineal transcripts. In the current study, we used higher resolution nonisotopic in situ hybridization techniques (Figure 7). Asmt transcripts were detectable in the cytoplasm of pineal parenchymal cells, whereas no signal was detected in the perivascular spaces or connective tissue septae of the gland (Figure 7A). The staining intensity was heterogeneous, varying considerably between pinealocytes: a subset of cells displayed a dense positive signal, whereas others were negative.

Figure 7.

In situ hybridization for detection of transcripts encoding melatonin-synthesizing enzymes in the rat pineal gland. A, Detection of Asmt transcripts by use of a DIG-labeled RNA probe. B, Detection of Tph1 transcripts by use of a DIG-labeled RNA probe. C, Detection of Aanat transcripts by use of a biotin-labeled DNA probe. The detected signal was pineal specific. The arrows indicate densely stained pinealocytes and the arrowheads indicate weakly stained pinealocytes. The asterisks indicate perivascular spaces. All animals were killed at ZT18. For specificity controls, see Supplemental Figure 4. Scale bar, 50 μm.

To determine whether other enzymes involved in melatonin synthesis exhibited a similar heterogeneous pattern of distribution, nonisotopic in situ hybridization studies were done with a probe for Tph1 (Figure 7B); TPH1 is the first enzyme in the synthesis of serotonin from tryptophan (Figure 1). In addition, sections were reacted with a probe for Aanat transcripts (Figure 7C); Aanat encodes the first enzyme in the conversion of serotonin to melatonin (Figure 1) and controls the daily rhythm in melatonin. Tph1 staining was clearly stronger than that of either Asmt or Aanat, the latter of which was the weaker. Both Tph1 and Aanat exhibited the markedly heterogeneous pattern seen with Asmt.

Immunohistochemical colocalization studies of ASMT and TPH1

The heterogeneous expression pattern of Tph1 revealed by in situ hybridization prompted us to investigate the distribution of TPH1 protein in the rat pineal gland and to colocalize TPH1 and ASMT (Figure 8). For this purpose, an antibody known to specifically detect tryptophan hydroxylase in paraformaldehyde-fixed tissue sections was used (19, 33). These studies confirmed the heterogeneous staining pattern of TPH1 in the pineal gland at the protein level (Figure 8B; Supplemental Figure 9); under the conditions we used to detect TPH1 in the pineal gland, TPH2 was detected in perikarya of the raphe nuclei (Supplemental Figure 9B), in which Tph2 transcripts and immunoreactivity have been reported to occur (25, 33). Other parts of the brain were negative (Supplemental Figure 9). Colocalization analyses in both the superficial (Figure 8, A and B) and deep (Figure 8, C and D) pineal gland revealed a subset of pinealocytes that was positively stained for both ASMT and TPH1; however, other pinealocytes were immunopositive for either one enzyme or the other.

Discussion

The rat pineal gland has been extensively used to study the regulation of pineal melatonin synthesis (10). However, due to the lack of reliable antisera and the small size of the pinealocyte, the melatonin-producing cell of the gland has not been histologically identified in tissue sections. The results of this investigation are of special interest with regard to several issues of importance to biology of the pineal gland. These include the heterogeneous nature of pinealocytes, sites of melatonin synthesis, and cellular localization of ASMT.

A biochemically heterogeneous nature of pinealocytes

The results of this investigation indicate that pineal parenchymal cells are highly heterogeneous with regard to indicators of melatonin synthesis. Pinealocytes comprise more than 95% of the cells in the gland (11), and this presumably accounts for its success in making apparently homogenous primary cell cultures for studying biochemical regulatory processes, eg, melatonin synthesis (10). Biochemical and molecular analyses normally generate data representing the sum of a large number of pinealocytes or whole glands, presumably masking the biology of the individual cell. However, our findings from immunohistochemical studies with ASMT, SAG, and TPH1 and from in situ hybridization localization of transcripts encoding TPH1, AANAT, and ASMT suggest a pronounced variation in the markers between individual cells and are consistent with prior observations of heterogeneity among pinealocytes, as detailed in the introductory text. The ASMT protein was detectable in a subset of pinealocytes. A localization of ASMT in solitary interstitial cells present in the pineal parenchyma cannot be excluded; however, because no interstitial cells in the perivascular spaces were immunopositive, we find it highly unlikely that melatonin synthesis occurs in this cell type. Thus, the results of our studies indicate that enzymes required for melatonin synthesis are present only in pinealocytes, consistent with the view that these cells are the source of melatonin.

However, it is not clear as to why staining intensity differs markedly between individual pinealocytes. Several explanations can be reasonably considered. One is that differential staining in the rat pineal gland reflects different biological states of individual pinealocytes of the same type. If this is the case, at any given time, only a subset of pinealocytes might contribute to melatonin synthesis and populations of cells would go through work/rest cycles in which all genes required for melatonin synthesis and associated processes would be coordinately controlled. However, the SAG/ASMT and TPH1/ASMT colocalization evidence from the current study indicates that expression of pineal marker genes is not coordinated and that marker genes are not expressed in a uniform pattern among pinealocytes. Accordingly, support for a work/rest cycle explanation is not strong.

It seems reasonable that the heterogeneous expression of pinealocyte marker genes, including those linked to melatonin synthesis, reflects the operation of different mechanisms controlling expression of these genes. In this scenario, multiple molecular mechanisms involving quantitatively and qualitatively different combinations of transcription factors, such as cone-rod homeobox, orthodenticle homeobox 2, retina and anterior neural fold homeobox, and nuclear receptor 4A1 (22, 28, 29, 34) and others (8), might act in parallel in different pinealocytes to establish different levels of melatonin synthesis genes. This points to the possibility that melatonin synthesis might involve multiple pinealocyte subtypes, each of which is primarily responsible for a specific step in melatonin synthesis. According to this production-line type of explanation, extracellular transit of intermediates plays an essential role, with serotonin being synthesized in one specialized cell type and then released and taken up by a second pinealocyte subtype to be converted to N-acetylserotonin (Figure 1). N-acetylserotonin would then be converted to melatonin, perhaps in the same cell type or a third. Support for the required extracellular transit of intermediates comes from observations that pineal glands release significant amounts of serotonin and N-acetylserotonin in addition to melatonin (35–40). Such a separation of function is consistent with the heterogeneous staining reported here and with the partial colocalization of ASMT and TPH1 in only a subset of pinealocytes.

Cellular localization of melatonin synthesis

Prior studies on the distribution of transcripts in the pineal gland have used radioisotopic methods. Whereas these methods are sensitive and very useful for the determination of the location of a transcript in brain regions and in neuronal perikarya (15–100 μm in diameter), this technique does not allow the detection in individual pinealocytes because of their small size (7–12 μm in diameter). The current study is the first to use in situ hybridization techniques capable of determining the cellular localization of Tph1, Aanat, and Asmt transcripts in the rat pineal gland. The results presented indicate that whereas Tph1 and Asmt transcripts are detectable using DIG-labeled riboprobes, a more sensitive biotin-based technique including amplification is required to detect Aanat transcripts. This is consistent with quantitative molecular determination of the abundance of these transcripts (8). This study extends the existing knowledge on tissue distribution from radiochemical in situ hybridization studies (3, 5, 6, 8, 25, 29, 32) to include the above-mentioned variation between transcript levels in individual pinealocytes of the pineal gland parenchyma.

At the subcellular level, rod-like structures containing both AANAT and ASMT, melatonin-producing complexes, have been reported in the cytoplasm of human pinealocytes, and in the same study, AANAT was also localized to similar rod-like structures in the sheep (16). ASMT is clearly localized to the cytoplasm in rat pinealocytes, which is an accord with previous reports on ASMT activity in the cytosol (41). However, we did not in the present study find the ASMT immunoreaction product to be localized to specific subcellular structures, suggesting that the previously reported complexes are species specific. Similarly, there are species differences in the regulation of pineal melatonin synthesis that also place rodents and nonrodent mammals in different groups: the melatonin rhythm in rodents depends on nocturnal de novo synthesis of Aanat transcripts and protein followed by cAMP-dependent phosphorylation of AANAT protein (10, 32). This leads to the stabilization and activation by association with 14-3-3 protein in a reversible complex. In sheep and primates, rhythmic melatonin synthesis seems to be solely driven by phosphorylation of AANAT translated from constitutively synthesized mRNA and association with 14-3-3 (10, 42).

Melatonin synthesis in the pineal complex

The finding of ASMT-positive pinealocytes in all parts of the pineal complex, including the stalk and the deep pineal gland, is in agreement with a similar distribution of Asmt transcripts in the rat pineal complex (6). Our finding that the heterogeneous pattern of ASMT-immunopositive cells is seen in all parts of the pineal complex provides added evidence for the similarity of the composition and organization of pineal complex elements. The similarity of all elements of the pineal complex revealed by ASMT immunodetection is consistent with evidence in the literature that there is a daily rhythm in melatonin synthesis in both the superficial and deep pineal gland that is generated by sympathetic innervation from the superior cervical ganglia (9, 43, 44). Accordingly, current and previous findings support the view that melatonin is produced in all parts of rat the pineal complex.

Extrapineal distribution of melatonin-synthesis

Previous analyses of ASMT activity have shown that this indicator is detected at high levels only in pineal tissue (45). This is in agreement with the Western blot analysis presented here. However, low levels of Asmt transcripts have been reported in a wide variety of extrapineal tissues in studies involving very sensitive PCR-based methods (46), and we cannot exclude the possibility of extrapineal presence of ASMT-containing cells as seen in our immunohistochemical analyses of the habenula. The mammalian neural retina is known to produce melatonin during nighttime (47). However, in the rat retina, the transcriptome of which shares a high degree of similarity with that of the pineal gland (8), Asmt expression levels 2 orders of magnitude below those of the pineal gland have been reported (5); in the data presented here, ASMT was below detectable levels in extracts of the whole retina, consistent with evidence of low expression and activity (5, 18, 41).

ASMT daily rhythm

As noted above, pineal melatonin synthesis exhibits a prominent circadian rhythm driven by oscillations in AANAT activity (1, 10). Pineal ASMT was detectable throughout the 24-hour daily cycle, and a significant diurnal rhythm was not detected. These findings are in accord with measurements of ASMT activity in the pineal gland, showing only very minor day-night differences (48, 49). It should be noted that a previous report of ASMT presence in the rat pineal gland during nighttime only was retracted (50, 51).

Use of ASMT and SAG as pinealocyte markers

SAG is widely regarded as a pinealocyte marker (12, 21); however, in contrast to ASMT, the function of SAG is unknown in the pineal gland. SAG is restricted to a subset of pinealocytes, as revealed in immunohistochemical studies of various mammalian species, including rat and man (12, 52). It has been speculated that the fact that only a fraction of pinealocytes are SAG positive may reflect the existence of two different pinealocyte subpopulations (12); however, as discussed above, SAG- and ASMT-positive pinealocytes do not represent the same subpopulation or mutually exclusive separate subpopulations of pineal cells, but rather two partially overlapping immunohistochemically defined populations of pinealocytes.

As with the distribution of ASMT, SAG is found in all parts of the pineal complex, and in the deep part, peripheral processes of the pinealocytes even penetrate the medial habenular nucleus (12, 31). SAG-positive cells, considered as pinealocyte-like based on both morphological criteria and the presence of SAG, have also been found in the habenula of various species (31, 53). Our data on ASMT colocalization and distribution in the rat suggests that these pinealocyte-like cells of the habenula are not just carrying a pinealocyte molecular identity in the form of S-antigen but may also contribute to melatonin synthesis.

SAG has been used to mark pinealocytes in anatomical studies (12, 31, 53), to identify pinealocytes in culture (54–56), and to identify pineal-derived tumors (57–59), but to our knowledge, it has not been used for colocalization purposes on tissue sections. Notably, the current study demonstrates the value of the monoclonal antibody PH8 for TPH colocalization investigations of the pineal gland. Using ASMT as a pinealocyte marker, also in immunohistochemical efforts with the purpose of colocalization, would seem appropriate in future studies. This is because of the nature of the immunoreaction, which in our hands makes it ideal to identify specific individual cells, and the direct involvement of ASMT in melatonin synthesis, a defining feature of the pinealocyte.

The results of this effort have established the utility of several tools in characterizing individual pinealocytes and hypothetical extrapineal sites of melatonin synthesis. This advance adds to the list of cytochemical approaches that can be used to characterize the cellular composition of the pineal gland, including pinealocytes, sympathetic nerve fibers, glial cells, and endothelial cells (Supplemental Table 1) (8, 12, 22, 60–68). With these tools it will be possible to better understand the cellular location of molecular processes and their relationships within the pineal gland, which will enhance the impact of biochemical approaches such as transcriptome and proteome profiling (8, 30, 69, 70) that cannot establish the cellular localization of specific molecules.