• Open Access

Serotonin receptors antagonistically modulate Caenorhabditis elegans longevity

Authors

  • Hana Murakami,

    1. Gheens Center on Aging, Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, Louisville, KY, USA
    Search for more papers by this author
  • Shin Murakami

    1. Gheens Center on Aging, Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, Louisville, KY, USA
    Search for more papers by this author


Shin Murakami, Gheens Center on Aging, Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, 580 S Preston Street, BaxterII, RM102, Louisville, KY 40202, USA. Tel.: 502-852-2554; fax: 502-852-2660; e-mail: shin.murakami@louisville.edu

Summary

The neurotransmitter serotonin has been implicated in affecting the variation of longevity in natural Drosophila populations and age-related diseases in mammals. Based on these observations, it has been predicted that serotonin signal, perhaps at levels of serotonin biosynthesis, may control lifespan. Here, we investigated a variety of mutations in serotonin-signal genes, including serotonin biosynthesis genes, a serotonin transporter gene, and serotonin receptor genes. Despite this prediction, mutations in the serotonin biosynthesis genes had little or modest effects on lifespan, while the mod-5 mutation with increased availability of serotonin caused a modest life-shortening effect. In contrast, a deletion mutation of the ser-1 serotonin receptor gene increased longevity by up to 46%, likely through the insulin/insulin-like growth factor 1 pathway. This result suggests an interaction between the serotonin pathway and the insulin/insulin-like growth factor 1 pathway. A deletion mutation of another serotonin receptor gene, ser-4, shortened early to mid lifespan. The results suggest that serotonin signal antagonistically modulates longevity through different serotonin receptors. This study may indicate serotonin receptors as a potential target for antigeric interventions.

Results and discussion

Serotonin (5-HT) is a dose-sensitive neurotransmitter that modulates a variety of biological processes and behaviors. In the fruit fly, a quantitative trait locus for the variation of longevity has been mapped in the aromatic l-amino acid decarboxylase (Ddc) gene, which is required for serotonin and dopamine biosynthesis (De Luca et al., 2003). In mammals, age-related changes in serotonin signal are known risk factors of age-related diseases, such as diabetes and cardiovascular diseases, in which serotonin receptors, 5-HT1A and 5-HT2C receptors, may play a role (Mattson et al., 2004). These observations in the fruit fly and in mammals prompted us to test the hypothesis that serotonin signal may modulate longevity in Caenorhabditis elegans.

We investigated mutations that alter levels of serotonin and/or dopamine. They include mutations in serotonin biosynthesis genes (tph-1 and bas-1; reducing serotonin), a mutation in a dopamine biosynthesis gene (cat-2; reducing dopamine), a double mutation in dopamine and serotonin biosynthesis genes (bas-1;cat-4; reducing both dopamine and serotonin), and a mutation in a serotonin transporter gene (mod-5; increasing serotonin availability) (Loer & Kenyon, 1993; Ranganathan et al., 2000, 2001; Sawin et al., 2000; Sze et al., 2000; Dempsey et al., 2005). The mod-5 mutation causes effects similar to serotonin-reuptake inhibitors that are used as antidepressants (Ranganathan et al., 2001). All the genes used in this study and their orthologues are shown in Table 1.

Table 1. Caenorhabditis elegans genes used in this study
GeneGene product
C. elegansFruit flyMouseHuman
  • The table includes the genes used, their orthologues and similar genes in other model species.

  • *

    The mod-1 gene is a unique serotonin gated ion channel, which is distantly related to GABA- and glycine-gated chloride channels (Ranganathan et al., 2000). An example of the genes similar to mod-1 is shown.

  • In the fruit fly, the serotonin receptor genes similar to ser-1 show similarity to two dopamine receptor genes, DopR2 (Dopamine receptor 2) and D2R (Dopamine 2 like receptor). Because it is known that SER-1 responds to serotonin but not to dopamine, the two Drosophila dopamine receptor genes were excluded from the table.

  • SER-1 and SER-4 have properties of both 5-HT1 and 5-HT2 receptors (see text).

bas-1DdcDdcDDCAromatic l-amino acid decarboxylase (DOPA decarboxylase)
cat-2PleThThTyrosine hydroxylase
cat-4PuGch1GCH1GTP cyclohydrolase I
daf-2InRInsr, Igf1r, InsrrINSR, IGF-1R, INSRRInsulin/IGF-1 receptor
daf-16foxo (dFOXO)Foxo1, Foxo3a, Mllt7 (Foxo4)FOXO1, FOXO3a, FOXO4FOXO/forkhead transcription factor
mod-1Rdl*Gabrb3*GABRB3*Unique serotonin-gated ion channel similar to mammalian GABA- and glycine-gated chloride channels
mod-5SerTSlc6a4SLC6A4Na+, Cl-dependent serotonin transporter
ser-15-HT1A, 5-HT2Htr2a, Htr2cHTR2A, HTR2C5-HT2 receptor
ser-45-HT1A, 5-HT1BHtr1a, Htr1bHTR1B, HTR1D, HTR1A, HTR1F5-HT1 receptor
tph-1TrhTph1, Tph2TPH1, TPH2Tryptophan hydroxylase

The results are summarized in Fig. 1 and Table 2. The serotonin biosynthesis mutations (tph-1 and bas-1) and the dopamine biosynthesis mutation (cat-2) caused no effect on lifespan (Fig. 1a,b). Thus, it is unlikely that reducing serotonin biosynthesis has an effect on lifespan under normal growth conditions. The tph-1 mutant had little effect on lifespan, but showed a prolonged reproduction period (P < 0.0001): Mean reproductive spans were: 4.4 ± 0.2 days for N2 (n = 23); 3.2 ± 0.0 days for ser-1(ok345) (n = 26); 5.1 ± 0.0 days for tph-1(mg280) (n= 17). The result of reproduction span in the tph-1 mutant is consistent with the previous study (Sze et al., 2000). The results suggest that there is an uncoupling between lifespan and reproductive lifespan, and it has been reported that there is such an uncoupling in many life-extending mutants (e.g. Murakami & Johnson, 1996; Gems et al., 1998; Jenkins et al., 2004). Interestingly, the tph-1 mutant modestly extended lifespan in the presence of FUDR (5-Fluoro-2′-deoxyuridine) (see below). In contrast, the mod-5 mutation caused shortened lifespan (P = 0.037; Fig. 1c). There was little difference in the survival curve of the mod-5 mutant during early mid phase of aging (day 2–8; before the end of reproduction). This result suggests that increasing the availability of serotonin may cause life-shortening. Taken together, reducing serotonin biosynthesis has little or modest effect on lifespan, while increasing serotonin may have a modest life-shortening effect. Note that the cat-2 mutant with reduced dopamine showed no effect on lifespan, while the bas-1; cat-4 double mutant with reduced dopamine and serotonin showed short lifespan (Fig. 1b).

Figure 1.

Lifespan of various serotonin signal mutants. Lifespan of (a) tph-1 and mod-1; (b) bas-1, cat-2, and bas-1; cat-2; (c) mod-5; and (d) ser-1 and ser-4. (e) and (f) Genetic interaction among the ser-1 mutation and the mutations in the insulin/IGF-1 pathway. Mean lifespans are shown in Supplementary Table 1. Asterisks indicate the strains fed with control RNAi bacteria.

Table 2.  Lifespans of various mutants
GenotypeMean lifespan (days)% changeMaximum lifespan (days)No. of replicationsnComment
  • Mean and SEM of two or more replications were shown. Probabilities were calculated by using the Gehan-Wilcoxon test.

  • *

    Strains were fed with the bacteria carrying control RNAi vector;

  • †Percentage of change in mean lifespan compared to control;

  • ‡Number of replications (sets) is shown;

  • §

    P < 0.0001,

  • P < 0.01.

Serotonin-related mutants
wild-type11.5 ± 0.3  –19.1 ± 1.38175 
bas-1(ad466)10.8 ± 0.2−6.117.3 ± 0.7490Serotonin biosynthesis
cat-2(e1112)11.4 ± 0.3−0.923.3 ± 2.1496Dopamine biosynthesis
bas-1(ad466); cat-4(e1141) 9.9 ± 0.9−14.014.3 ± 0.7490Dopamine and serotonin biosynthesis
mod-1(ok103)11.9 ± 0.33.525.2 ± 0.24101Serotonin-gated chloride channel
mod-5(n822)10.7 ± 0.3−7.019.5 ± 2.3489Serotonin reuptake
ser-1(ok345)15.4 ± 0.4§34.027.0 ± 1.2488Serotonin receptor
ser-1(ok345) + p(ser-1)12.0 ± 0.44.322.0 ± 1.7244 
ser-4(ok512)10.2 ± 0.3−11.321.3 ± 2.1490Serotonin receptor P = 0.01 (day0–15)
tph-1(mg280)11.4 ± 0.4−0.923.3 ± 1.5499Serotonin biosynthesis
Genetic epistasis1
wild-type*12.0 ± 0.4  –17.6 ± 0.34100 
ser-1(ok345)*14.9 ± 0.4§24.222.5 ± 0.94100 
daf-2(RNAi)14.3 ± 0.4§19.222.5 ± 0.94100Insulin/IGF-1 receptor
ser-1(ok345); daf-2(RNAi)14.8 ± 0.4§23.322.0 ± 1.04100 
Genetic epistasis2
wild-type*11.3 ± 0.3  –17.4 ± 0.28179 
ser-1(ok345)*15.9 ± 0.5§40.721.8 ± 0.68181 
daf-16(RNAi)11.1 ± 0.2−1.814.5 ± 0.48163Forkhead transcription factor
ser-1(ok345); daf-16(RNAi)11.5 ± 0.21.814.6 ± 0.48179 
Genetic epistasis3
wild-type*12.0 ± 0.5§  –20.0 ± 1.1479 
eat-2(ad465)*16.0 ± 0.7§33.324.0 ± 1.7479 
ser-1(RNAi)17.5 ± 0.5§45.829.3 ± 1.1484 
ser-1(RNAi); eat-2(ad465)20.9 ± 1.1§74.238.5 ± 3.1481 
mod-5(n822)* 9.5 ± 0.3§−20.817.8 ± 0.3480 
ser-1(RNAi); mod-5(n822)12.2 ± 0.11.722.0 ± 1.2481 
FUDR/serotonin treated
Treated with 1 mg mL−1 serotonin
wild-type-FUDR15.5 ± 0.4  –22.8 ± 0.8479 
wild-type-FUDR + serotonin14.7 ± 0.5−5.220.0 ± 0.7481 
tph-1(mg280)-FUDR18.6 ± 0.420.028.8 ± 1.4471 
tph-1(mg280)-FUDR + serotonin17.1 ± 0.310.322.8 ± 0.9468 
ser-1(ok345)-FUDR17.7 ± 1.2§14.231.8 ± 0.6474 
ser-1(ok345)-FUDR + serotonin16.1 ± 0.53.926.5 ± 0.3474 
Treated with 5 mg mL−1 serotonin
wild-type-FUDR14.6 ± 0.7  –18.0 ± 0.6477 
wild-type-FUDR + serotonin14.8 ± 1.01.418.5 ± 0.5479 
tph-1(mg280)-FUDR16.7 ± 0.4§14.124.0 ± 0.0474 
tph-1(mg280)-FUDR + serotonin13.6 ± 0.7−6.820.5 ± 0.5477 
ser-1(ok345)-FUDR16.8 ± 1.6§15.124.5 ± 1.7479 
ser-1(ok345)-FUDR + serotonin17.8 ± 0.9§21.924.0 ± 0.0461 

We further investigated the hypothesis that serotonin influences lifespan at levels of downstream receptors. To this end, we tested mutants in three serotonin receptor genes, including two seven-transmembrane G-protein-coupled receptors for serotonin (ser-1 and ser-4) and a serotonin-gated chloride channel gene (mod-1) (Table 1). It has been shown that SER-1 responds to serotonin but not to other biogenic amines, including dopamine, histamine, and tryptamine (Hadman et al., 1999). SER-1 has properties of both 5-HT1 receptor (5-HT1R) and 5-HT2 receptor (5-HT2R). By the pharmacological definition, SER-1 is 5-HT1R (Hadman et al., 1999; Dempsey et al., 2005), while by the protein similarity, SER-1 is closely related to 5-HT2R (Hadman et al., 1999; http://www.wormbase.org). Similarly, SER-4 has the pharmacological property of 5-HT1R and is structurally similar to 5-HT2R (Demsey et al., 2005; WormBase; http://www.wormbase.org). Interestingly, a BLAST search identified the fruit fly dopamine receptors (DopR2 and D2R) and serotonin receptors (5-HT1A and 5-HT2) that are structurally similar to SER-1 (http://www.wormbase.org; data not shown). Because SER-1 has an activity as a serotonin receptor but not as a dopamine receptor (Hadman et al., 1999), it is likely that the dopamine receptor genes are structurally similar but functionally different. Note that the MOD-1 serotonin chloride channel is unique to C. elegans, and, thus, an example of similar genes is shown in Table 1.

The three serotonin-receptor mutations had different effects on lifespan. A deletion mutation in ser-1 significantly extended mean lifespan by up to 46% and maximum lifespan by 41% (Table 2). The life extension of the ser-1 mutant was suppressed by expression of the ser-1 plasmid (Table 2). While the mod-1 mutation had no effect on lifespan (Fig. 1a), a deletion mutation in ser-4 shortened early to mid lifespan (from 0 to 15 days after hatching; P = 0.01); the life-shortening effect of ser-4 was not significant during overall lifespan (P = 0.07). There was no clear abnormality in morphology observed in the ser-4 animals (data not shown). We suggest that serotonin receptors, ser-1 and ser-4, antagonistically affect lifespan.

The life-extension phenotype of the ser-1 mutant was largely suppressed by daf-16(RNAi) (Fig. 1e; Table 2). In addition, the ser-1 mutant did not increase extended lifespan of the daf-2(RNAi) animals. The life extension by ser-1(RNAi) was additive to that of the eat-2 mutant, which causes calorie restriction (Fig. 2a). It has been shown that there was no significant difference between the rate of pharyngeal pumping in the ser-1 animals and the wild-type animals, although there is a slight increase in variability in the rate (Hobson et al., 2006). Thus, the results suggest: (i) the ser-1 serotonin signal interlocks into the insulin/insulin-like growth factor 1 (IGF-1) pathway; (ii) the ser-1 life extension is independent of calorie restriction. It is yet to be determined how the serotonin pathway interacts with the insulin/IGF-1 pathway. Note that the long-lived ser-1(RNAi) animals did not show altered expression of an insulin-like peptide gene marker, daf-28::GFP (Supplementary Fig. 1). Similarly, ser-4(RNAi) did not affect expression of daf-28::GFP (data not shown). We also tested the effect of the mod-5 mutation that reduces removal of serotonin. Although the mod-5 mutation reduced the lifespan of the ser-1 mutation (Fig. 2b), it is likely that reduced lifespan of the mod-5 mutation causes this result.

Figure 2.

Effects of the eat-2 mutant, FUDR and serotonin on lifespan in the ser-1 mutant. (a) Genetic interaction between the eat-2 mutation and the ser-1 mutation; (b) genetic interaction between the mod-5 mutation and the ser-1 mutation; (c) the effects of FUDR and serotonin on lifespan of the tph-1 animals; (d) the effects of FUDR and serotonin (5 mg mL−1) on lifespan of the ser-1 animals. Mean lifespans are shown in Supplementary Table 1. Asterisks indicate the strains fed with control RNAi bacteria.

Interestingly, the tph-1 mutant showed life extension by 20% in the presence of FUDR (Fig. 2c) but not under normal growth conditions (Fig. 1a). FUDR is known to suppress DNA synthesis, thus knocking out proliferating cells in reproductive tissues, eggs, and larvae. It is unlikely that internal hatching shortened lifespan in tph-1 under normal conditions, because we did not observe any contribution of such internal hatching to life-shortening. The life extension by the tph-1 mutation but not by the ser-1 mutation was suppressed by the supplementation of serotonin (P < 0.002; Fig. 2c,d; Table 2). The tph-1 and ser-1 mutations conferred resistance to heat and ultraviolet (UV) (Fig. 3).

Figure 3.

Stress resistance of the ser-1 and tph-1 mutants. (a) Thermotolerance of the tph-1 and mod-1 mutants; (b) Ultraviolet (UV) resistance of the mutants. Animals at the age of day 3 were incubated at 35 °C or exposed to UV light, and then fraction surviving was monitored. Thermotolerance: the tph-1(mg280) and ser-1(ok345) animals were significantly more resistant to heat than the wild-type (P < 0.0001). Mean LD50 was: for tph-1(mg280), > 600 min (n = 74); for ser-1(ok345), > 600 min (n = 80); for N2, 508 ± 17 min (n = 80). UV resistance: the tph-1(mg280) and ser-1(ok345) animals were also significantly more resistant to UV light than the wild-type (P < 0.0001). Mean LD50 was; for tph-1(mg280), 2.6 ± 0.2 days (N = 74); for ser-1(ok345), 2.5 ± 0.2 days (N = 80); for N2, 1.9 ± 0.1 days (n = 80).

In this study, we tested the hypothesis that serotonin signal influences lifespan based on observations in the fruit fly and in the mouse (De Luca et al., 2003; Mattson et al., 2004). Surprisingly, the major effects on lifespan were not observed in the mutants of the serotonin biosynthesis genes, but were observed in the mutants of the two serotonin receptors. The ser-1 and ser-4 mutants showed opposing effects on lifespan. This finding suggests a model in which altering levels of serotonin could trigger both life-extending and life-shortening effects depending on the serotonin receptors (Supplementary Fig. 2). In this model, reducing serotonin by the tph-1 mutations triggers the opposing effects on lifespan, suggesting that the tph-1 mutations do not necessarily cause the life-extension phenotype similar to ser-1 (ok345) and ser-1(RNAi). Note that the daf-16 nuclear localization is affected by the tph-1 mutant in response to starvation (Liang et al., 2006). This model also does not contradict the result that the mod-5 mutation caused a modest life-shortening effect. Interestingly, the ser-1 serotonin receptor gene is an example of the genes that have been missed in the genome-wide RNAi (RNA-mediated gene inactivation) screenings for lifespan (Dillin et al., 2002; Lee et al., 2003; Hamilton et al., 2005; Hansen et al., 2005). In fact, the rate of false negatives in an RNAi screening is relatively high (10–30%) even within the same laboratory (Simmer et al., 2003). This study suggests serotonin receptors as potential therapeutic targets for the prevention of age-related diseases and for anti-aging interventions.

Experimental procedures

Strain and media

General worm maintenance was as described in Murakami & Murakami (2005) and Murakami et al. (2005). Worms were grown at 25 °C on Nematode Growth Media (NGM) agar plates seeded with food, OP50 Escherichia coli. The genotypes of the strains used were: bas-1(ad446) III, mod-1(ok103) V, mod-5(n822) I, bas-1(ad446) III, cat-4(e1141) V, ser-1(ok345) X, ser-4(ok512) III, and tph-1(mg280) II. To avoid background variation, each strain was backcrossed three times or more. In our conditions, we did not see dauer formation in tph-1(mg280) and ser-1(ok345), although there were slow growing animals in tph-1(mg280) at a low frequency.

Assessment of survivorship, stress resistance, and reproductive span

Lifespan was assessed at 25 °C on NGM plates seeded with E. coli, OP50, as a food source as described in Murakami & Johnson (2001). Two sets of 20–25 worms were placed on NGM agar plates and transferred every day during the reproduction period, then transferred every 2 to 3 days until dead. The number of live worms was recorded for each plate, and the dead worms were noted and removed from the plate. For the FUDR and serotonin treatment, young adults at the age of day 2 were assessed for lifespan on the plates supplemented 0.05 mg mL−1 of FUDR and/or serotonin (1 mm or 5 mm) where noted. Live worms were transferred to a new plate with fresh bacteria every few days. Thermotolerance assay was performed as described previously (Murakami & Johnson, 2001). UV resistance assay was performed as described previously (Murakami & Johnson, 2001) with a modification. Young adults were irradiated by UV light (254 nm, 2.0 J cm−2) at the rate of 5.625 J m−2 per second. Reproductive spans were monitored by the period in which worms laid eggs (Murakami et al., 2005). Data were recorded and analyzed using Wilcoxon (Gehan) survival statistic implemented in the NCSS statistical software (NCSS Inc., Kaysville, UT, USA). LD50 was calculated by using the probit analysis in the NCSS software. All the experiments were performed more than twice.

RNA-mediated gene inactivation (RNAi)

daf-2(RNAi) and daf-16(RNAi) were performed by feeding RNAi bacteria on NGM plates supplemented with carbenicillin (Murakami et al. in press). RNAi bacteria carrying a control vector, daf-2(RNAi) vector, and daf-16(RNAi) were gifts from Dr Andrew Dillin (Salk Institute, CA, USA) (Dillin et al., 2002). The E. coli strain carrying the ser-1(RNAi) plasmid was from a C. elegans RNAi library (ORFeome version 1.1, Open Biosystems Inc., Huntsville, AL, USA) and verified by DNA sequencing. The RNAi bacteria were cultured overnight at 37 °C and spotted on the NGM plates with carbenicillin. RNA expression was induced by adding 100 µL of the isopropyl β-D-1-thiogalactopyranoside (IPTG)/carbenicillin solution (0.1 m IPTG, 25 µg mL−1 carbenicillin), followed by incubation at room temperature for up to 2 h. To ensure RNAi during early embryo, worms at the L3/L4 larval stage were grown with RNAi bacteria and were allowed to lay eggs. The eggs from the worms fed with each RNAi bacteria were placed on NGM plates seeded with the same RNAi bacteria. The worms were assessed for lifespan assay. The RNAi worms were fed with each RNAi throughout lifespan.

Acknowledgments

We thank Mr Jason Hellmann and Mr Danh Do for technical assistance, and the Caenorhabditis Genetics Center for strains and Dr Andrew Dillin for the RNAi plasmids. This work was supported by the Gheens foundation and the University of Louisville.

Ancillary