Disruption of a Neuropeptide Gene, flp-1, Causes Multiple Behavioral Defects in Caenorhabditis elegans

See allHide authors and affiliations

Science  11 Sep 1998:
Vol. 281, Issue 5383, pp. 1686-1690
DOI: 10.1126/science.281.5383.1686


Neuropeptides serve as important signaling molecules in the nervous system. The FMRFamide (Phe-Met-Arg-Phe-amide)–related neuropeptide gene family in the nematode Caenorhabditis elegans is composed of at least 18 genes that may encode 53 distinct FMRFamide-related peptides. Disruption of one of these genes,flp-1, causes numerous behavioral defects, including uncoordination, hyperactivity, and insensitivity to high osmolarity. Conversely, overexpression of flp-1 results in the reciprocal phenotypes. On the basis of epistasis analysis,flp-1 gene products appear to signal upstream of a G protein–coupled second messenger system. These results demonstrate that varying the levels of FLP-1 neuropeptides can profoundly affect behavior and that members of this large neuropeptide gene family are not functionally redundant in C. elegans.

FMRFamide-related neuropeptides (FaRPs) represent a large family of peptides that have been implicated as neurotransmitters or neuromodulators in many invertebrate and vertebrate behaviors, including muscular control (1), cardioregulation (2), pain modulation (3), and learning (4). Although a single organism may express multiple FaRPs encoded by multiple genes (5), little is known about the function of distinct FaRPs or the cellular and molecular mechanisms whereby different FaRPs exert their effects. We used a genetic approach to determine the role of a distinct set of FaRPs in the simple nematode model system, Caenorhabditis elegans.

Eighteen genes, designated flp-1 (FMRFamide-like peptide) through flp-18, in C. elegans potentially encode 53 distinct FaRPs, all of which contain the COOH-terminal RFamide (Arg-Phe-NH2) epitope characteristic of the FaRP family. At least 15 of these flp genes are expressed (6,7). The flp-1 gene consists of six exons and produces two transcripts by alternative splicing; these two transcripts encode seven distinct FaRPs, which range from seven to ten amino acids in length and which contain a FLRFamide sequence at their COOH-termini (6). By standard two- and three-factor crosses, we determined that flp-1 is located on the right arm of chromosome IV between unc-24 and dpy-20 (Fig. 1A) (8). This map position has been confirmed by the C. elegans Genome Sequencing Consortium (9), which places flp-1 on cosmid F23B2. To determine the expression pattern of flp-1, transgenic animals carrying lacZ or green fluorescent protein (GFP) reporter constructs were generated. Thus far, we have detected flp-1 expression only in neurons in the anterior region of the animal (Fig. 1D) (10).flp-1–expressing cells include AVK, AVA, AVE, RIG, RMG, AIY, AIA, and M5. These cells are a subset of the FMRFamide-like immunoreactive cells previously described (11).

Figure 1

flp-1 deletion alleles. (A) Genetic position of flp-1 on the right arm of chromosome IV. (B) Genomic organization of flp-1. Top line: partial restriction map of the flp-1 genomic region. E, Eco RI; C, Cla I. Bottom line: Expanded genomic region to show organization of flp-1. Exons are indicated as boxes (peptide coding sequences are indicated in black), introns as lines; the site of the Tc1 insertion is indicated by an inverted triangle. Two deletions of the flp-1 gene were isolated by PCR screening of populations of Tc1 insertion strains for an imprecise Tc1 excision (13). The extent of the yn2(1.4 kbp) and yn4 (2.1 kbp) deletions are shown; both deletions remove upstream and coding regions. Primers used to screen for deletions are indicated by arrows. (C) Backcrossed yn2 and yn4 strains are homozygous deletion lines. Genomic DNA was isolated from backcrossedyn2 and yn4 animals, digested with Cla I, and Southern blotted. The position of the probe is shown in (B). Solid arrow indicates wild-type hybridizing fragment; double arrows indicate smaller hybridizing fragments resulting from the yn2 andyn4 deletions. No hybridizing fragment corresponding to the wild-type fragment is detected in DNA from homozygous yn2and yn4 animals. Molecular weight markers (kilobase pairs) are indicated to the left. (D) Expression offlp-1 in wild-type animals is detected only in anterior head neurons. Transgenic animals were stained for β-galactosidase activity to visualize expression of flp-1-lacZ reporter constructs (10). Scale bar, 25 μm.

To determine the function of flp-1 in C. elegans, we screened for flp-1 deletion mutants by using a Tc1 transposon-dependent polymerase chain reaction (PCR) method (12). Populations of a Tc1 transposon insertional allele offlp-1 [NL242 (pk41)] were screened for germline deletions (13). Mutants with 1.4 kilobase pairs (kbp) [NY7(yn2)] and 2.1 kbp [NY16 (yn4)] deleted from the flp-1 gene were isolated (Fig. 1B). By DNA sequence analysis, we determined that the yn2 deletion removed 570 bp of the promoter sequence, the start site of transcription (6), exons 1 through 3, and most of exon 4; theyn4 deletion removed 1.1 kbp of the promoter sequence, the start site of transcription, and exons 1 through 4. Thus, both alleles are likely to represent severe loss of flp-1 function. DNA was isolated for Southern blot analysis to confirm the presence of the deletions in homozygous strains (Fig. 1C) (14). The pattern of FMRFamide-like immunoreactivity did not differ betweenflp-1 mutants and wild-type animals (15), suggesting that individual cells expressing FLP-1 peptides also express other FaRPs.

To determine how loss of FLP-1 peptides might affect behavior, bothflp-1 mutant strains were backcrossed multiple times into a wild-type Bristol background to remove unlinked copies of Tc1. Several motor defects were immediately apparent in flp-1 animals. Wild-type animals move on a food source with a characteristic sinusoidal waveform (16). By contrast, homozygousflp-1 mutants had loopy, uncoordinated movement; the sinusoidal waveform was exaggerated (Fig. 2), such that the wave amplitude was significantly greater than that of the wild type (Table 1). The action of FLP-1 peptides on waveform may be through the flp-1–expressing interneurons, AVE and AVA, both of which synapse onto ventral nerve cord motoneurons important for coordinated movement in the animal (17).

Figure 2

Varying the level of flp-1expression has reciprocal effects on locomotion. (A) Wild-type animal shows a characteristic sinusoidal waveform; sinusoidal tracks can also be seen in the bacterial lawn. (B)flp-1(yn2) animal. flp-1(yn2) animals show an exaggerated waveform characterized by an increased wave amplitude.flp-1(yn4) animals had a less loopy phenotype thanflp-1(yn2) animals (Table 1). (C) Animal in which flp-1 is overexpressed (ynIs9). After heat shock, transgenic animals carrying a flp-1 cDNA under the control of a heat shock promoter show a much flattened waveform compared to that of wild-type animals (Table 1). Scale bar, 0.1 mm.

Table 1

Phenotypes seen in flp-1 deletion and overexpression animals. Adult hermaphrodites (except where noted) were scored for nose touch, osmotic avoidance, uncoordination, movement rate, and wandering. Mean values ± standard errors are reported. Numbers of animals tested are indicated in parentheses. ND, not determined; NC, not calculated.

View this table:

A second obvious motor abnormality seen in flp-1 animals was hyperactivity, characterized by an increased rate of movement and a tendency to crawl off the agar onto the side of the plate even in the presence of food (wandering). To quantitate hyperactivity, we counted the number of body bends per minute on a food source and recorded the frequency with which the animals crawled off the agar. Homozygousyn2 animals were significantly more hyperactive than wild-type animals and showed significant wandering behavior (Table 1). Surprisingly, homozygous yn4 animals, which have the larger deletion in flp-1, were not as hyperactive as yn2animals and showed little wandering behavior (Table 1). This difference in behavior between the two strains suggests that there is a low level of expression from the remaining flp-1 exons inyn4 animals, perhaps due to expression from a cryptic promoter, as has been reported for other C. elegans genes with deletions in the 5′ region (18). Alternatively, a novel deleterious product may be made in yn2 animals, resulting in a more severe phenotype. Trans-heterozygotes with one yn2and one yn4 allele, however, show phenotypes (hyperactivity, wandering, and uncoordination) similar to yn4 homozygotes (15), suggesting that the more severe phenotypes in theyn2 animals are not due to a novel product. Hyperactive and wandering behaviors are also seen in food-deprived animals (19). Because flp-1 animals have normal chemotaxis responses to soluble and volatile substances (including food), flp-1 animals may be defective in the pathway responsible for signaling animals to slow down once they reach food.

Several sensory behaviors were also examined in the flp-1deletion animals. Wild-type animals recoil when lightly touched on the nose (nose touch sensitivity) (20), move backward or forward when gently touched on the head or tail, respectively (mechanosensitivity) (21), avoid areas of high osmolarity (osmotic avoidance) (22), and are attracted or repelled by specific soluble and volatile chemicals (chemosensitivity) (23); for the most part, these four behaviors are mediated by independent neural circuits. Two sensory defects were detected inflp-1 animals (Table 1). First, flp-1 mutants were nose touch–insensitive: homozygous flp-1 animals responded to nose touch stimuli less than 30% of the time compared to ∼80% of the time for wild-type animals. Second, homozygousflp-1 animals were defective for osmotic avoidance and did not hesitate to cross a high osmolarity barrier. FLP-1 peptides, therefore, are necessary in the pathways for response to nose touch and high osmolarity. Two flp-1–expressing interneurons, AVA and AIA, receive direct synaptic input from the ASH sensory neurons, which are multi-modal sensory neurons that transduce both nose touch and osmotic stimuli (20). flp-1 mutants had wild-type body mechano- and chemosensory responses (24), indicating that FLP-1 peptides are not necessary for these behaviors.

The entire coding region of flp-1 lies within the intron of a putative gene encoded on the opposite DNA strand (9); because the yn2 and yn4 deletions extend further than the flp-1 coding region, a small part of the coding region of this putative gene is removed. To confirm that the phenotypes seen in flp-1 mutants are the result of the absence of FLP-1 peptides, we performed germline transformation offlp-1 animals with a plasmid containing the wild-typeflp-1 gene and 6 kbp of flp-1 upstream sequence (25). flp-1 transgenic animals carrying the plasmid (n = 37) were rescued for coordination (67% rescued), hyperactivity (70% rescued), and wandering (94% rescued), confirming that control of these behaviors is dependent on FLP-1 peptides.

To determine the effect of overexpression of FLP-1 peptides, we generated transgenic animals carrying a flp-1 cDNA under the control of a heat shock promoter, hsp 16-2 (26,27), in a wild-type (ynIs9 and ynIs10) or aflp-1 (yn2;ynIs9) mutant background (28). When ynIs9 or ynIs10 adult animals were heat-shocked for 1 to 2 hours at 33°C, their movement was decreased and animals showed a flattened waveform compared to that of wild-type animals that had been similarly heat-shocked (Fig. 2 and Table 1). Furthermore, although 80% of wild-type (n = 90) and flp-1(n = 41) animals continued to move after heat shock, only 32% of the flp-1 transgenic animals (n= 56) were moving after heat shock, and this movement was highly sluggish. Overexpression of FLP-1 peptides, therefore, results in reciprocal movement phenotypes to those seen in flp-1reduction-of-function animals, further demonstrating that FLP-1 peptides are involved in controlling the rate of movement and muscle coordination. Interestingly, yn2;ynIs9 animals, which contained the heat shock transgene in a flp-1 deletion background, showed slightly more severe movement phenotypes after heat shock than those seen in heat-shocked ynIs9 orynIs10 animals, which contained the transgene in a wild-type background. The waveform of heat-shocked yn2;ynIs9 animals disappeared and basically corresponded to the width of the animal (Table 1); moreover, less yn2;ynIs9 animals (22%;n = 27) were moving after heat shock thanyn9 or yn10 animals (32%). Recovery from the effects of flp-1 overexpression occurred after about 18 hours (n = 24). The increased severity of the behaviors in a flp-1 deletion background may be due to a compensatory up-regulation of the peptide receptors, resulting in greater sensitivity when the peptides are overexpressed. Because of the extreme sluggishness of the transgenic animals after heat shock, we were unable to test the animals for osmotic avoidance. However,yn2;ynIs9 animals responded to nose-touch whenflp-1 was overexpressed; moreover, this sensitivity was reversible, as the animals became nose touch–insensitive again after overnight recovery (Table 1).

Serotonin (5-hydroxytryptamine or 5-HT) also regulates the rate of movement in C. elegans. Exogenous 5-HT has an inhibitory effect on movement in C. elegans (29), whereas 5-HT–deficient animals are hyperactive (30). To determine whether FLP-1 peptides and 5-HT are involved in a common pathway controlling movement, we tested whether flp-1 animals are sensitive to exogenous 5-HT. When exposed to 6.5 mM 5-HT, the movement of wild-type animals was significantly inhibited (Table 2). By contrast, the movement offlp-1 animals was not significantly decreased by 5-HT (Table 2). FLP-1 peptides, therefore, appear to act downstream or parallel to serotonin in this signaling pathway. For instance, theflp-1–expressing interneurons AVA, AVE, or AIA could be influenced by the serotonergic NSM (29) to modulate locomotion.

Table 2

flp-1 deletion animals are resistant to serotonin-induced movement inhibition. Adult animals were placed on agar plates with or without 6.5 mM serotonin for 2 hours. Movement rates in the presence (#) or absence (∧) of food (absence of food increases movement rates of animals) were determined as in Table 1.

View this table:

Most FaRPs are likely to signal through a heterotrimeric GTP–binding protein (G protein)–coupled receptor (31); to date, however, the only FaRP receptor isolated is a FMRFamide-gated sodium channel (32). Caenorhabditis elegans animals with a mutation in a G protein subunit, such as Gαo (goa-1), Gαq(egl-30), or Gβ (gpb-1), show multiple phenotypes, including uncoordination and movement rate phenotypes similar to those in flp-1 mutants (30,33), suggesting that FLP-1 peptides act via a G protein–coupled second messenger system. To test this hypothesis, we generated double mutants between flp-1 and G protein–subunit mutant animals as follows: a flp-1 heat shock overexpression animal (ynIs9) was mated with a Gαo loss-of-function [goa-1(n363)] (30) or a Gαq gain-of-function [egl-30(syEx125)] mutant animal (33), and aflp-1(yn2) reduction-of-function mutant animal was mated with a constitutively active Gαo[goa-1(syIs9)] (33), a loss-of-function Gαq [egl-30(ad809)] (33), or a Gβ overexpression [gpb-1(pkIs372)] (33) mutant animal. In each case, the resulting double mutant demonstrated uncoordination and movement rate phenotypes similar to those of the single G protein–subunit mutant (Table 3). These results suggest that FLP-1 peptides signal upstream of these G proteins to control coordinated movement. Recent work has shown that Gαq is likely to act downstream of Gαo (34); FLP-1 may be one of the signals that activate the G protein pathway.

Table 3

FLP-1 peptides signal upstream of a G protein–mediated pathway. Double mutants with an flp-1allele and an allele of one of the G protein subunits were generated; the phenotype of adult animals was scored as in Table 1.

View this table:

Our work demonstrates that individual FaRP-encoding genes have unique functions in an organism. FLP-1 peptides are critical in the functioning of multiple sensory and locomotory pathways and act as neuromodulators of the classical neurotransmitter serotonin in movement inhibition. Varying the levels of FLP-1 peptides results in a spectrum of behavioral phenotypes in C. elegans, as demonstrated by the range of phenotypes seen in the reduction-of-function animals,yn2 and yn4, and the flp-1-induced overexpression animals. Seven distinct FaRPs are encoded byflp-1; it is unknown whether expression of any FLP-1 peptide is sufficient for normal sensory and motor function or whether certain peptides have unique sensory or motor functions. However, FLP-1 peptides have functions distinct from other FaRPs in C. elegans. Given the number of additional flp genes and the possibility that each gene or peptide has specific, nonredundant functions, the possible complexity of FaRP activity in C. elegans is enormous. Caenorhabditis elegans will provide a valuable system in which the function of a complex neuropeptide gene family can be elucidated.

  • * Present address: Axys Pharmaceuticals, 180 Kimball Way, South San Francisco, CA 94080, USA.

  • Present address: Elan Pharmaceuticals, 3760 Haven Avenue, Menlo Park, CA 94025, USA.

  • To whom correspondence should be addressed. E-mail: li{at}


View Abstract

Navigate This Article