Genetics, Vol. 148, 703-718, February 1998, Copyright © 1998, Genetics Society of America

An Insulin-like Signaling Pathway Affects Both Longevity and Reproduction in Caenorhabditis elegans

Heidi A. Tissenbauma and Gary Ruvkuna
a Department of Molecular Biology, Massachusetts General Hospital, and Department of Genetics, Harvard Medical School, Boston, Massachusetts 02114

Corresponding author: Gary Ruvkun, Department of Molecular Biology, Massachusetts General Hospital, Wellman Building--8th floor, 50 Blossom St., Boston, MA 02114, ruvkun{at}molbio.mgh.harvard.edu (E-mail).

Communicating editor: I. GREENWALD


*  ABSTRACT
*TOP
 *ABSTRACT
*MATERIALS AND METHODS
 *RESULTS
 *DISCUSSION
 *LITERATURE CITED

Mutations in daf-2 and age-1 cause a dramatic increase in longevity as well as developmental arrest at the dauer diapause stage in Caenorhabditis elegans. daf-2 and age-1 encode components of an insulin-like signaling pathway. Both daf-2 and age-1 act at a similar point in the genetic epistasis pathway for dauer arrest and longevity and regulate the activity of the daf-16 gene. Mutations in daf-16 cause a dauer-defective phenotype and are epistatic to the diapause arrest and life span extension phenotypes of daf-2 and age-1 mutants. Here we show that mutations in this pathway also affect fertility and embryonic development. Weak daf-2 alleles, and maternally rescued age-1 alleles that cause life span extension but do not arrest at the dauer stage, also reduce fertility and viability. We find that age-1(hx546) has reduced both maternal and zygotic age-1 activity. daf-16 mutations suppress all of the daf-2 and age-1 phenotypes, including dauer arrest, life span extension, reduced fertility, and viability defects. These data show that insulin signaling, mediated by DAF-2 through the AGE-1 phosphatidylinositol-3-OH kinase, regulates reproduction and embryonic development, as well as dauer diapause and life span, and that DAF-16 transduces these signals. The regulation of fertility, life span, and metabolism by an insulin-like signaling pathway is similar to the endocrine regulation of metabolism and fertility by mammalian insulin signaling.

GENETIC screens for mutants with increased life spans have identified genes that may regulate the aging process (GUARENTE 1996 Down; KENYON 1996 Down). In many cases, the mutations are pleiotropic, affecting life span in conjunction with altered metabolism, stress resistance, or reproduction (FRIEDMAN and JOHNSON 1988A Down; KENNEDY et al. 1995 Down; VANFLETEREN and DE VREESE 1995 Down). An inverse correlation between life span and metabolism rate has been noted in phylogenetic comparisons of life span (FINCH 1990 Down). Additionally, regimens that alter metabolism such as caloric restriction extend the maximum life span in rats and mice (FINCH 1990 Down; MASORO et al. 1991 Down). In the nematode Caenorhabditis elegans two genes known to regulate aging, daf-2 and age-1, also affect the rate of metabolism (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; KENYON et al. 1993 Down; DORMAN et al. 1995 Down; LARSEN et al. 1995 Down; MORRIS et al. 1996 Down; KIMURA et al. 1997 Down). As well, mutations in these genes result in developmental arrest and metabolic shift at the dauer larval stage (GOTTLIEB and RUVKUN 1994 Down; LARSEN et al. 1995 Down; MALONE et al. 1996 Down; KIMURA et al. 1997 Down) (reviewed in RIDDLE 1988 Down; RIDDLE and ALBERT 1997 Down).

The C. elegans dauer larva is a specialized third larval stage adapted for survival in nonoptimal environmental conditions. In response to high levels of a continuously secreted pheromone and low amounts of food, that is, unfavorable growth conditions, animals form a dauer larva (GOLDEN and RIDDLE 1984 Down). The dauer larva is metabolically shifted and stress resistant (CASSADA and RUSSELL 1975 Down; reviewed in RIDDLE 1988 Down; THOMAS 1993 Down; RIDDLE and ALBERT 1997 Down; KIMURA et al. 1997 Down). Animals arrested at the dauer stage can live up to eight times as long as a nondauer. Further, the time an animal spends as a dauer does not affect the post-dauer life span (KLASS and HIRSCH 1976 Down). When conditions become more favorable, animals molt and reenter the life cycle at the fourth larval stage (L4). These recovered animals are nearly indistinguishable from animals that have not arrested at the dauer stage (RIDDLE and ALBERT 1997 Down; reviewed in RIDDLE 1988 Down).

Genes that affect dauer formation (daf) fall into two classes: genes that mutate to a dauer-constitutive phenotype, where animals enter dauer inappropriately, and dauer-defective mutants, where, even under unfavorable growth conditions, animals will not arrest as a dauer. Based on genetic epistasis analysis, these genes have been ordered into a pathway (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; THOMAS et al. 1993 Down; GOTTLIEB and RUVKUN 1994 Down). daf-2 and age-1 are placed in a similar branch of the dauer pathway since they show unique and similar epistasis behavior (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; KENYON et al. 1993 Down; GOTTLIEB and RUVKUN 1994 Down; MALONE and THOMAS 1994 Down; DORMAN et al. 1995 Down; LARSEN et al. 1995 Down).

Unlike all other daf genes, weak or conditional alleles of daf-2 and age-1 exhibit an increase in the post-dauer life span (KENYON et al. 1993 Down; DORMAN et al. 1995 Down; LARSEN et al. 1995 Down; MORRIS et al. 1996 Down). Therefore, these daf-2 and age-1 mutants affect life span without the associated developmental arrest at the dauer stage, suggesting that dauer arrest can be decoupled from life span extension in these animals. Both daf-2 and age-1 have been shown by genetic epistasis analysis to regulate daf-16 activity: mutations in daf-16 suppress the dauer arrest and longevity phenotypes of both daf-2 and age-1 (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; KENYON et al. 1993 Down; GOTTLIEB and RUVKUN 1994 Down; DORMAN et al. 1995 Down).

Molecular characterization of daf-2 and age-1 indicates that an insulin-like signaling pathway regulates diapause and longevity in C. elegans: daf-2 encodes a member of the insulin receptor family and age-1 encodes the catalytic subunit of phosphatidylinositol-3-kinase (PI-3-kinase), a molecule that has been shown to act downstream of the mammalian insulin receptor (reviewed in KAHN 1994 Down; MORRIS et al. 1996 Down; KIMURA et al. 1997 Down). Insulin plays a central role in regulating mammalian metabolism. Not only is daf-2 related to the mammalian insulin receptor by sequence homology, but it also regulates metabolism: decreased DAF-2 signaling induces metabolic changes (KIMURA et al. 1997 Down).

In a genetic screen for long-lived mutants, age-1 (hx546) was recovered (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down). age-1(hx546) fails to maternally complement age-1(mg44), which is associated with a molecular lesion in a PI 3-kinase gene (MORRIS et al. 1996 Down). Moreover, genetic mapping places age-1(hx546) in the approximate region of the PI 3-kinase gene affected by the age-1 alleles mg44, mg109, and m333 (GOTTLIEB and RUVKUN 1994 Down; MALONE et al. 1996 Down; MORRIS et al. 1996 Down). However, sequencing of the PI 3-kinase coding region of age-1 (hx546) failed to reveal a mutation (MORRIS et al. 1996 Down). Therefore, there is some doubt whether age-1(hx546) affects this PI-3-kinase gene or identifies a closely linked maternal enhancer mutation. In this study, we further mapped age-1(hx546) to refine its genetic map position to the same interval as the PI-3-kinase gene. We also show that age-1(hx546) affects both maternal and zygotic age-1 activity and that putative null mutations in the PI 3-kinase gene enhance age-1(hx546). We find that complete lack of zygotic and maternal age-1 leads to dauer arrest, whereas lack of maternal or reduction of zygotic age-1 activity leads to an increase in life span in the absence of dauer arrest. These data suggest that normal senescence depends on phosphatidylinositol signaling from both maternal and zygotic AGE-1.

We show that at the nonpermissive temperature, daf-2 mutants show markedly reduced fertility as well as partially penetrant embryonic lethal phenotypes. Similarly, we show that maternally rescued age-1 null mutants cause the same spectrum of phenotypes. The increased life span and dauer constitutive phenotypes (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; KENYON et al. 1993 Down; GOTTLIEB and RUVKUN 1994 Down; DORMAN et al. 1995 Down) as well as the decreased fertility and embryonic lethal phenotypes of daf-2 and age-1 mutants are suppressed by daf-16 mutations. Thus, the daf-2 and age-1 regulation of life span and reproduction uses the same pathway as the dauer arrest pathway. We argue that both the aging and reproductive effects of daf-2 and age-1 mutations derive from the metabolic changes induced by insulin-like signaling defects.


*  MATERIALS AND METHODS
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
*RESULTS
 *DISCUSSION
 *LITERATURE CITED

Methods and strains:
All strains used were maintained and handled as described in BRENNER 1974 Down and SULSTON and HODGKIN 1988 Down. Animals were grown on standard NG agar plates supplemented with 4 µM streptomycin and 10 µM nystatin to minimize mold. In this study, the mutations used were LGI: dpy-5(e61), daf-16(m27, mg51, mg52, mg53, and mg54); LGII: unc-4(e120), sqt-1(sc13), age-1(hx546, mg44, mg109, and m333), lin-29(n333), mnC1; LGIII: daf-7(e1372), daf-2(e1370, e1391), dpy-17(e164), unc-32(e189); LGIV: unc-5(e53); LGV: dpy-11(e229), osm-3(p802); LGX: daf-12(m20); lon-2(e678).

Mapping between unc-4 and sqt-1:
age-1(hx546) was three-factor mapped in the 1.4 map unit (Acedb) unc-4 sqt-1 interval. age-1(hx546) males were obtained by heat shock and mated into unc-4(e120) sqt-1(sc13) hermaphrodites at 20°. Non-Sqt-non-Unc progeny were picked to separate plates. From these plates, recombinants (Sqt-non-Unc and Unc-non-Sqt animals) were picked to individual plates and allowed to self-fertilize to isolate a homozygous strain.

Mapping between sqt-1 and lin-29 :
age-1(hx546) was three-factor mapped in the 1.2 map unit (GOTTLIEB and RUVKUN 1994 Down) sqt-1 lin-29 interval. age-1(hx546) males were obtained by heat shock and mated with sqt-1(sc13) lin-29(n333)/mnC1 hermaphrodites at 20°. Non-Sqt-non-Lin progeny were picked to separate plates. From plates that segregated Sqt Lin progeny, recombinants (Sqt-non-Lin and Lin-non-Sqt animals) were picked to individual plates and allowed to self-fertilize to isolate a homozygous strain.

Complementation test:
age-1(hx546) males were mated with sqt-1(sc13) age-1(mg44) hermaphrodites at 20°. F1 Non-Sqt Non-Daf cross progeny were picked to separate plates. Individual heterozygotes were singled to plates at 25°, 20°, or 15° and transferred daily (20° and 25°) or every 2–3 days (20° and 15°). Progeny of the age-1(hx546)/sqt-1(sc13) age-1(mg44) strain were counted and scored for dauer and non-dauer after either 3 days (25°), 4 days (20°) or 7 days (15°). For sqt-1(sc13) age-1(mg44)/sqt-1(sc13)age-1(mg44) progeny from sqt-1(sc13) age-1(mg44)/sqt-1(sc13)age-1(mg44) mothers, hermaphrodites were allowed to lay eggs for 6-14 hours and then removed from the plate, because these animals tend to be egg-laying defective such that their eggs hatched within them. Plates were scored for dauer and nondauer after either 3 days (25°), 4 days (20°) or 7 days (15°).

Testing recombinants for maternal age-1 activity:
unc-4(e120) age-1(m333)/mnC1 males were mated with all of the recombinant hermaphrodites with the exception outlined below. Wild-type hermaphrodite cross-progeny were singled to 25° and their progeny scored for whether unc-4(e120) age-1 (m333) homozygous progeny arrest as dauer larvae. For the Unc-non-Sqt recombinants, sqt-1(sc13) age-1(mg44)/mnC1 males were obtained by heat shock and were mated with the recombinants. For the Lin-non-Sqt recombinants, sqt-1(sc13) age-1 (mg44)/mnC1; him-8(e1489) males were mated with the homozygous Lin recombinants. Prior to mating, Lin recombinants were opened at the vulva with an injection needle.

Isolation of new daf-16 alleles:
daf-2(e1370); daf-12(m20) animals were grown at 15° and then placed in a bleach solution to isolate eggs. Animals were then grown until the L4 and mutagenized with ethylmethanesulfonate, then placed at 15° and allowed to grow for one full generation. Gravid F1s were grouped together and then placed in a bleach solution to isolate F2 eggs. The eggs were placed in S medium overnight at 25° on a rocking platform. Each 15 ml tube contained a different set of animals and remained separate. The synchronous preparations of F2 larvae were placed on large plates seeded with bacteria at 25° and 2–3 days later examined for any animals that failed to arrest at the dauer stage. Each individual plate from independent mutagenized parents on which a nondauer suppressor mutant was isolated was named mg51, mg52, mg53, mg54. Only one suppressor mutant was studied from each egg pool; thus the suppressor mutants are independent.

Mapping and complementation tests of mg51, mg52, mg53, mg54:
Because daf-16 had been previously identified as a suppressor of daf-2 (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; GOTTLIEB and RUVKUN 1994 Down), mg51, mg52, mg53, and mg54 were first tested for complementation of daf-16(m27) as follows: daf-16(m27); daf-2(e1370) males were mated into each of the mutant strains at 25°. Two and three days later the mating plate was examined for the presence of dauers and nondauers.

The newly isolated mutants were also genetically mapped. At 15°, daf-2(e1370) males were mated with mg51, daf-2(e1370), daf-12(m20) hermaphrodites and male cross-progeny were picked. These mg51/+; daf-2(e1370); daf-12(m20) males were mated with each of the following hermaphrodite strains; Chromosome I dpy-5(e61); daf-2(e1370); Chromosome II sqt-1 (sc13); daf-2(e1370); Chromosome III daf-2(e1370) unc-32 (e189); Chromosome IV daf-2(e1370); unc-5(e53); Chromosome V daf-2(e1370); dpy-11(e229); Chromosome X daf-2 (e1370); lon-2(e678). The cross progeny were either daf-2 (e1370); mg51/+; marker/+; daf-12(m20)/+ or daf-2(e1370); marker/mg51; daf-12(m20)/+. From these plates, nonmarked individual progeny were singled at 15° and allowed to self-fertilize. From this brood, a total of 50 L4s or young adults homozygous marker progeny were singled to 25° from 2–3 individual plates and their broods examined for dauers and nondauers. If the marker was unlinked to the suppressor, then the progeny were either marked dauers, marked arrested larvae (daf-2; daf-12), or marked dauers and nondauers. If the marker was linked to the suppressor then all of the animals would be dauers, unless there was recombination between the marker and the suppressor.

Life span assays:
Life span assays were done at 25°. Adult hermaphrodites were picked (5–6 per plate) from each strain and allowed to undergo one full generation at 15° or 20° to ensure the animals being tested had not starved or gone through dauer. From these plates, individual L4 or young adult animals were picked to plates at 25°. Occasionally, individual L4 or young adult animals were picked from a well-seeded nonstarved plate as well. Day 1 of the life span was the day that the animal was picked to 25°. Therefore, life spans are either post-L4 or post-young adult. Animals were transferred to new plates every 1–2 days while producing progeny. After egg production ceased, animals were transferred to new plates every 4–7 days. Animals were tapped with a pick every 2–4 days and were scored as dead when they did not move after repeated taps with the pick. Animals that did not have any progeny, were egg-laying defective such that their eggs hatched within them, or had crawled off the plate, were not included in the study. For each analysis, N2 and age-1 (hx546) were also tested for life span as controls, to account for any changes in the incubator environment.

Fertility measures:
Adult hermaphrodites were picked (5–6 per plate) from each strain and allowed to undergo one full generation at 15° or 20° to ensure the parental strain had not starved or gone through dauer. From these plates, individual L4 or young adult animals were singled to individual plates at 25°. The parental animals were transferred daily to fresh plates and the number of eggs laid overnight was counted. Plates at 25° were scored for dauers, nondauers, dead eggs and arrested larvae two days later. Dauers that had crawled off the side of the plate were counted as well. Plates at 15° were scored for dauers, nondauers, dead eggs, and arrested larvae 5–6 days later.

Strain constructions:
daf-16(m27); daf-2(e1391): daf-16(m27) males were mated into daf-2(e1391) hermaphrodites at 15°. Cross-progeny were singled to individual plates at 25° and allowed to self-fertilize. The F2 brood plates segregated dauers and nondauers. From one plate, sixteen dauers were singled to 15° and allowed to recover. When the F2 animals developed into fertile adults, they were shifted again to 25° and allowed to self-fertilize. The progeny were then scored for dauers and nondauers. Many nondauers were singled onto separate plates from a plate segregating both dauers and nondauers [daf-16(m27)/+; daf-2(e1391) parent]. A non-dauer that produced all nondauers at 25° and established the strain was tested to confirm the presence of daf-2 in the strain by crossing to daf-2(e1370) males and looking for dauer cross-progeny in the F1 at 25°.

daf-16(m27); daf-2(e1370); osm-5(p802):
daf-16(m27); daf-2(e1370) males were mated with daf-2(e1370); osm-5(p805) hermaphrodites at 15°. Cross-progeny were singled to individual plates at 25° and allowed to self-fertilize. The F2 brood plates segregated dauers and nondauers. However, since the daf-2(e1370); osm-5(p805) strain had been previously reported to result in 50–75% arrested L1s at 25° (VOWELS and THOMAS 1992 Down), cross-progeny plates that lacked arrested L1s and had nondauers were picked for further analysis. Of the eight cross-progeny plates, two plates had nondauers and few arrested larvae. From these plates, 48 animals were singled to individual plates and allowed to self-fertilize. Each individual plate was then tested for the presence of osm-5(p802) by testing for the ability to take up 3-3'-dioctadecyloxacarbocyanine perchlorate (DiO) as described in PERKINS et al. 1986 Down.


*  RESULTS
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
 *RESULTS
*DISCUSSION
 *LITERATURE CITED

Fine mapping of age-1(hx546):
Animals carrying the age-1(hx546) allele live 2–3 times as long as wild-type animals, and arrest at the dauer stage at temperatures above which animals are standardly cultivated (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; DORMAN et al. 1995 Down; MALONE et al. 1996 Down; MORRIS et al. 1996 Down). The strong age-1 alleles, mg44, mg109, and m333, are associated with amino acid substitutions or nonsense mutations in the PI-3-kinase gene that maps between sqt-1 and lin-29 (GOTTLIEB and RUVKUN 1994 Down; LARSEN et al. 1995 Down; MORRIS et al. 1996 Down; Figure 1; Table 1). age-1(hx546) fails to maternally complement, but zygotically complements molecular null mutations in the PI-3-kinase (MORRIS et al. 1996 Down; Figure 1; data not shown). age-1 (hx546) has also been mapped to the sqt-1 lin-29 interval where the PI-3-kinase gene is located (MALONE et al. 1996 Down; MORRIS et al. 1996 Down). However, no PI-3-kinase coding sequence change has been detected in age-1(hx546) (MORRIS et al. 1996 Down). Thus, the genetic evidence that age-1(hx546) affects the PI-3-kinase gene activity has not been confirmed by molecular analysis. To further substantiate that age-1(hx546) affects the same PI-3-kinase gene affected by mg44, mg109, and m333, we studied genetic interactions between these known PI-3-kinase mutations and age-1(hx546). We also mapped age-1(hx546) further to delimit its genetic map position to the PI-3-kinase genetic region.




View larger version (76K):
In this window
In a new window
Download PPT slide
  		Figure 1. age-1(hx546) fails to provide maternal age-1 activity at all temperatures. In this experiment, age-1(mg44) was marked in cis with sqt-1(sc13). The + chromosome in this figure is the balancer mnC1. age-1(hx546) males were mated with sqt-1(sc13) age-1(mg44) hermaphrodites at 20°. F1 Non-Sqt Non-Daf cross-progeny were picked to separate plates. Individual heterozygotes were singled to plates at either 25°, 20° or 15°. Animals were then counted and scored for dauer and nondauer as outlined in MATERIALS AND METHODS. In this figure, progeny of age-1(mg44)/age-1(mg44) parents are on the left, progeny of age-1(mg44)/age-1(hx546) parents are in the middle, and progeny of age-1(mg44)/+ parents are on the right. In each data set, the left column represents data collected at 25°, the middle column is data at 20°, and the right column is data from animals grown at 15°. At 25°, 94% of the animals arrested as dauers, and 6% of the animals arrested as L1s or dead eggs; from age-1(mg44)/age-1(hx546) parents, all Sqt animals arrested as dauers except one fertile adult was a recombinant sqt-1(sc13) age-1(mg44)/sqt-1(sc13) age-1(hx546); from age-1(mg44)/+ parents, all fertile age-1(mg44)/age-1(mg44) adults gave rise to all Sqt dauers in the F3 except for three recombinants sqt-1(sc13) age-1(mg44)/sqt-1(sc13). At 20°, from age-1(mg44)/age-1(hx546) parents, all fertile age-1(mg44)/age-1(mg44) adults gave rise to all Sqt dauers in the F3 except for one recombinant sqt-1(sc13) age-1(age-1(mg44)/sqt-1(sc13) age-1(hx546); from age-1(mg44)/+ parents, all fertile age-1(mg44)/age-1(mg44) adults gave rise to all Sqt dauers in the F3 except for one recombinant sqt-1(sc13) age-1 mg44)/sqt-1(sc13). At 15°, from age-1(mg44)/age-1(hx546) parents; from age-1(mg44)/+ parents all fertile age-1(mg44)/age-1(mg44) adults gave rise to all Sqt dauers in the F3 except for one recombinant sqt-1(sc13) age-1(age-1(mg44)/sqt-1(sc13) age-1(hx546); from age-1(mg44)/1, all fertile age-1(mg44)/age-1(mg44) adults gave rise to all Sqt dauers in the F3 except for one recombinant sqt-1(sc13) age-1(mg44)/sqt-1(sc13). The genotype of the recombinants was determined by picking individual animals from the original nondauer Sqt and then examining its brood. The complementation tests were performed as outlined in MATERIALS AND METHODS.


 
View this table:
In this window
In a new window

  		Table 1. Three factor map of age-1(hx546)

age-1(mg44), a putative null allele of age-1, substitutes a stop codon at amino acid position 405 (Trp405Amber) that is predicted to truncate AGE-1 upstream of the kinase domain (MORRIS et al. 1996 Down). Animals with only maternally contributed age-1 activity, age-1(mg44)/age-1(mg44) progeny of age-1(mg44)/+ mothers live twice as long as wild type animals but do not arrest at the dauer stage (MORRIS et al. 1996 Down; Figure 1). However, animals bearing no maternal or zygotic age-1 activity [progeny of age-1(mg44)/age-1(mg44) long-lived mothers] arrest at the dauer stage [GOTTLIEB and RUVKUN 1994; MORRIS et al. 1996; Figure 1; in all the experiments reported below, age-1(mg44) is marked in cis with sqt-1(sc13)]. Therefore, the putative null phenotype of age-1 is maternal-effect dauer-constitutive and zygotic long-lived (GOTTLIEB and RUVKUN 1994 Down; MORRIS et al. 1996 Down; Figure 1).

age-1(hx546) was originally reported to map to the left of sqt-1, close to unc-4 (FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down). Our data conflicts with this mapping. We collected recombinants between age-1(hx546) in trans to an unc-4 sqt-1 chromosome. All of the Sqt-non-Unc (14/14) recombinants were non-Age; animals bearing each of these recombinant chromosomes maternally complemented age-1(mg44) (Table 1). However, all of the Unc-non-Sqt (3/3) recombinants were Age: they failed to supply maternal age-1 activity. This is consistent with age-1(hx546) mapping to the right of sqt-1 or within 0.29 map units (P < 0.05) to the left of sqt-1.

Recombinants were also tested for life span (Table 1; all of the life span values are mean ± SE). 3/3 Sqt-non-Unc non-Age recombinants that maternally complemented age-1(m333) mutants had life spans similar to wild-type (8.9 ± 0.2 days, n = 170). The mean life spans of the recombinants were 9.7 ± 0.7 (n = 25), 9.7 ± 0.4 (n = 31), and 8.1 ± 0.4 days (n = 29). An Unc-non-Sqt Age recombinant that failed to complement age-1(mg44) maternally was long-lived relative to wild type with a mean life span of 18.5 ± 1.8 days (n = 21). Animals that carry the marker mutations sqt-1(sc13) or unc-4(e120) had mean life spans similar to wild type, of 11.4 ± 0.4 (n = 34) and 10.8 ± 1.3 days (n = 13), respectively.

In summary, we found a total of 17 recombinants in the unc-4 sqt-1 interval. All of the Unc-non-Sqt recombinants fail to maternally complement age-1(mg44) and a recombinant tested for life span was long-lived. The Sqt-non-Unc recombinants maternally complemented age-1(m333) and the three recombinants tested for life span were short-lived (Table 1).

We also collected recombinants between a sqt-1 lin-29 chromosome in trans to age-1(hx546). For the 17 Sqt-non-Lin recombinants, 3/17 were Sqt-Age and 14/17 were Sqt-non-Age (MORRIS et al. 1996 Down; Table 1). Of the 6 Lin-non-Sqt recombinants, there were 4/6 Lin-Age recombinants and 2/6 Lin-non-Age recombinants (MORRIS et al. 1996 Down; Table 1). The Lin-non-Age non-Sqt recombinants, as well as the other mapping in Table 1, help to eliminate the possibility that the age-1(hx546) mutation maps to the left of sqt-1, as previously reported (FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down). Of the total 23 recombinants collected in the sqt-1 lin-29 interval, 5 were between sqt-1 and age-1 and 18 were between age-1 and lin-29 (MORRIS et al. 1996 Down; Table 1). In addition, the age-1(hx546) failure to maternally rescue age-1(mg44) and the long life span are linked in these 35 recombinants over 2.3 map units. Therefore, all of the genetic mapping data is consistent with the age-1(hx546) lesion lying in the PI-3-kinase gene product. It is possible that the age-1(hx546) mutation is in a region that has not been sequenced, such as in a promoter or enhancer element (QUIRING et al. 1994 Down; ORTIZ-LOPEZ et al. 1997 Down). Sequencing of the age-1 flanking regions in age-1(hx546) may identify such an element.

age-1(hx546) affects maternal age-1 gene activity at all temperatures:
age-1(hx546) fails to contribute maternal age-1 activity at 25°: 100% of the age-1(mg44) homozygous progeny of age-1(hx546)/age-1(mg44) heterozygotes arrest at the dauer stage (MORRIS et al. 1996 Down; Figure 1). This is the same as the control age-1(mg44) daughters of age-1(mg44) parents (GOTTLIEB and RUVKUN 1994 Down; MORRIS et al. 1996 Down; Figure 1). The dauer larvae produced in both cases have the dark intestine, cuticular remodeling, and arrest of the molting cycle characteristic of this stage (reviewed in RIDDLE 1988 Down; RIDDLE and ALBERT 1997 Down; GOTTLIEB and RUVKUN 1994 Down). In contrast, only 1% of age-1(mg44) progeny of +/age-1(mg44) parents arrest development as dauer larvae at 25° (Figure 1).

age-1(hx546) supplies sufficient zygotic activity to allow reproductive development: age-1(hx546)/age-1 (mg44) progeny of age-1(hx546)/age-1(mg44) mothers form fertile adults, like +/age-1(mg44) progeny (MORRIS et al. 1996 Down; Figure 1). In addition, in the absence of age-1 maternal contribution, zygotic age-1 expression from age-1(hx546) is sufficient to allow reproductive development: mating age-1(hx546) males with age-1(mg44) hermaphrodites, the cross-progeny age-1 (hx546)/age-1(mg44) develop as reproductive adults rather than dauers. Therefore, at 25° age-1(hx546) expresses sufficient zygotic but not maternal levels of age-1 for nondauer growth.

age-1(hx546) supplies some age-1 maternal gene activity at lower temperatures (Figure 1). At 20°, 100% of age-1(mg44) progeny from age-1(mg44) hermaphrodites arrest as dauers. However, fewer of the age-1(mg44) progeny of age-1(hx546)/age-1(mg44) arrest development as dauer larvae at 20°; most continue development as dauer-like animals that are dark, developing into sterile adults, and 4% of age-1(mg44) progeny of age-1(hx546)/age-1(mg44) parents are fertile (Figure 1). Therefore, age-1(hx546) maternal activity is temperature sensitive.

At 15°, age-1(hx546) shows considerable maternal age-1 activity, although less than wild-type levels. Fifty-five percent of age-1(mg44) progeny of age-1(hx546)/age-1(mg44) parents are fertile (Figure 1). Because the age-1(mg44)/age-1(mg44) progeny of age-1(hx546)/age-1(mg44) hetero-zygous parent do not all arrest as dauers at lower temperatures, the age-1(hx546) mutant has sufficient maternal age-1 gene activity at lower temperatures to develop as a reproductive adult. There is still, however, a decrease in maternal age-1(hx546) activity: 45% of age-1(mg44) progeny of age-1(hx546)/age-1(mg44) parents arrest development without reproduction. However, most of the arrested animals develop to sterile adults rather than dauers. This suggests that decreases in maternal age-1 activity cause developmental arrest as a sterile adult, whereas complete lack of both zygotic and maternal age-1 activity results in arrest as a dauer larva.

The data suggest that age-1(hx546) is a temperature-sensitive allele. Alternatively, it is possible that the reduced level of maternal age-1(hx546) activity is sufficient at lower temperatures. This difference is not due to the intrinsic temperature sensitivity of the dauer pathway because, at all these temperatures, 100% of the progeny of parents carrying null age-1 alleles arrest as dauers (GOTTLIEB and RUVKUN 1994 Down; MORRIS et al. 1996 Down; Figure 1). Similar results were observed at all three temperatures when another marker, unc-4, was used to mark two other age-1 alleles, m333, and mg109 (data not shown). Both mg44 and m333 are molecular null alleles as they have premature stop codons, while mg109 is a missense mutation in a conserved region just outside the kinase domain (MORRIS et al. 1996 Down).

Decreased zygotic age-1(hx546) activity confers long life:
Animals lacking both zygotic and maternal age-1 activity will arrest development as dauers [age-1(mg44)/age-1(mg44) progeny from age-1(mg44)/age-1(mg44) mothers]. Lack of only maternal age-1 activity does not affect life span or development; age-1(mg44)/+ animals derived from age-1(mg44)/age-1(mg44) mothers develop as nondauers with normal senescence (Table 2 and Table 3; Figure 2A). Thus, age-1 is completely zygotically rescued (GOTTLIEB and RUVKUN 1994 Down; MORRIS et al. 1996 Down; Table 2 and Table 3; Figure 2A). Animals lacking only zygotic age-1 activity [age-1(mg44)/age-1(mg44) progeny of age-1(mg44)/+ mothers] develop into fertile adults but show a more than two-fold increase in life span (GOTTLIEB and RUVKUN 1994 Down; MORRIS et al. 1996 Down; Table 2 and Table 3; Figure 2A). These data indicate that reducing either maternal or zygotic age-1 activity by half has no effect on longevity or development as measured in our assays (Table 2 and Table 3; Figure 2, A and B).




View larger version (54K):
In this window
In a new window
Download PPT slide
  		Figure 2. (A) age-1(hx546) affects zygotic age-1 activity for life span. The mean life span and the number of animals examined are listed in Table 3. age-1(hx546)/age-1(mg44) hermaphrodites do not arrest at the dauer stage but the animals have an increased life span. In contrast, age-1(mg44)/+ progeny of age-1(mg44)/age-1(mg44) have a life span that is similar to +/+. Therefore, age-1(hx546) is defective in zygotic age-1 activity. Life span assays were performed as outlined in MATERIALS AND METHODS. Animals were grown at 20° until the L4 stage, when they were transferred to 25°. (B) age-1(mg44) enhances age-1(hx546) both maternally and zygotically. age-1(mg44)/age-1(hx546) progeny from age-1(mg44)/age-1(hx546) mothers have similar life spans to age-1(hx546)/age-1(mg44) hermaphrodite progeny of age-1(mg44)/age-1(mg44) mothers mated with age-1(hx546) males. This suggests that the age-1(mg44)/age-1(hx546) heterozygote may completely lack age-1 activity for life span. These data also suggest that age-1(hx546) has reduced but non-null zygotic gene activity for life span and age-1(mg44) enhances age-1(hx546) both maternally and zygotically. The small difference in age-1(hx546)/age-1(hx546) hermaphrodites from age-1(mg44)/age-1(hx546) when compared to the age-1(hx546) strain itself is not significant (P > 0.05). hx546/hx546* represents a backcrossed strain. Life span assays were performed as outlined in MATERIALS AND METHODS. Animals were grown at 20° until the L4 stage, when they were transferred to 25°. The mean life span for each strain and the total number of animals examined is shown in Table 3.


 
View this table:
In this window
In a new window

  		Table 2. The affect of maternal and zygotic age-1 genotype on dauer formation and longevity


 
View this table:
In this window
In a new window

  		Table 3. age-1 (hx546) has reduced maternal and zygotic age-1 activity

age-1(hx546) animals show an increase in life span but do not arrest at the dauer stage under normal growth conditions (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; DORMAN et al. 1995 Down; MALONE et al. 1996 Down; MORRIS et al. 1996 Down). Like other age-1 alleles, age-1(hx546) is zygotically rescued for life span: age-1(hx546)/+ animals from age-1(hx546)/age-1(hx546) mothers have wild-type life spans (Table 2 and Table 3; Figure 2A). age-1(hx546), however, does affect zygotic age-1 activity in the control of senescence. age-1(hx546) in trans to a putative null allele of age-1 supplies sufficient zygotic activity in the absence of any maternal activity to allow reproductive development in age-1(hx546)/age-1(mg44) progeny of age-1(mg44)/age-1(mg44) mothers, but these animals show a more than two-fold increase in longevity (Table 2 and Table 3; Figure 2, A and B). In fact, age-1(hx546) is zygotically enhanced for life span extension by age-1(mg44) (Figure 2, A and B and Table 2 and Table 3): age-1(hx546) has a mean life span of 15.6 ± 0.7 days while age-1(hx546)/age-1(mg44) animals have a mean life span of 19.8 ± 1.1 days. However, this is not due to the fact that age-1(mg44) is dominant: age-1(mg44)/+ progeny of age-1(mg44)/age-1(mg44) mothers show reproductive development and normal life span (Table 2 Table 3 Table 4; Figure 2A).


 
View this table:
In this window
In a new window

  		Table 4. Decreased brood size and viability in long-lived daf-2 and age-1 mutants

These results suggest that age-1(hx546) is a maternal enhancer of age-1(mg44). In addition, age-1(mg44) is a zygotic enhancer of age-1(hx546). These data along with the genetic mapping results presented in Table 1 and Figure 1 further endorse that age-1(hx546) is a mutation in the PI-3-kinase.

daf-2 and age-1 have pleiotropic effects on reproduction and viability:
daf-2(e1370) and daf-2(e1391) are temperature-sensitive dauer-constitutive alleles that substitute conserved amino acids in the kinase domain of the DAF-2 insulin-like receptor protein (KIMURA et al. 1997 Down). When shifted to the nonpermissive temperature after the larval stage 1 temperature-sensitive period (tsp) for dauer formation (SWANSON and RIDDLE 1981 Down), several daf-2 mutant animals are sterile (data not shown), while the remaining animals have a severely reduced fertility (Table 4). Wild-type animals grown at 20° until the young adult stage, and then shifted to 25°, had a mean brood size of 278 ± 20.4 with no embryonic lethality. However, daf-2(e1370) animals, shifted from 15° to the nonpermissive temperature (25°) at the young adult stage, had a severely reduced mean brood size of 67 ± 8.2 (Table 4). Moreover, 8.9% of the eggs laid resulted in dead eggs or arrested L1/L2 larvae. Most of the L1/L2 larvae recovered to form fertile adult hermaphrodites when shifted to 15°.

At the permissive temperature, daf-2(e1370) and daf-2(e1391) animals produce nearly wild-type brood size (Table 5). As well, less than 1% of the progeny of these daf-2 mutants arrest at embryonic and larval stages. Thus, only at the nonpermissive temperature do the daf-2 mutant alleles significantly reduce fertility and viability of progeny.


 
View this table:
In this window
In a new window

  		Table 5. daf-2 mutants have wild-type brood sizes and reduced lethality at 15°

We further examined whether or not the decreased fertility was due to a mutation unrelated to daf-2. We constructed a daf-2(e1370) dpy-17(e164) double mutant to eliminate any closely linked as well as unlinked mutations. This strain was then crossed to wild-type and a Daf-non-Dpy recombinant was collected in order to isolate a twice backcrossed homozygous daf-2 strain. This resulted in a slight increase in brood size and a slight decrease in the associated larval arrest/dead eggs (Table 4 and Table 5). However, the resulting backcrossed daf-2 (e1370) strain still exhibited a reduction in the brood size with associated larval arrest/dead eggs when shifted either as young adult (88 ± 8 total progeny with 6.9% arrest/dead eggs) or as L4 animals (71 ± 7 total progeny with 9.3% arrest/dead eggs). Because the reduction in fertility, the arrested larvae and dead eggs are seen in two daf-2 alleles as well as in a backcrossed strain; these data suggest that daf-2 controls function in other than dauer formation and that the reduction in brood size is indeed a phenotype of the daf-2 mutation itself.

Because age-1 and daf-2 show many similar defects, we examined the fertility of age-1 at 25°. age-1(mg44) homozygous animals also have severely reduced fertility at 25° with a mean of 76 total progeny (Table 4). Moreover, 5.8% of the total brood result in either dead eggs or larval arrest similar to daf-2 mutants. This reduction in fertility is not specific to this allele. age-1(hx546) also shows a 26% reduction in fertility compared to wild-type N2 at 25° (FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; Table 4). Therefore, insulin-like signaling through the AGE-1 PI-3-kinase gene also regulates reproduction, embryonic development, and dauer formation.

The amount of time that daf-2 mutants are allowed to develop at the permissive temperature affects fertility. As shown in Table 4, daf-2(e1370) mutants shifted to 25° at L4 have a mean brood size of 44 with 15.4% of the total brood either arrested as L1/L2 larvae or dying as eggs. Whereas, if daf-2(e1370) animals are shifted as young adults, the brood size increases to a mean of 67 with 8.9% arrest/lethality. A similar trend is shown for both daf-2(e1391) and the daf-2(e1370) backcrossed strain (Table 4). This indicates that daf-2 acts late in development to affect reproduction and progeny viability and, therefore, has activity well past the temperature-sensitive period for dauer formation.

In the life span assays, we noted the days that the animals laid progeny. In general, wild-type N2 animals produced progeny for the first four days of its 8-day life span. Therefore, approximately half of the wild-type animal's adult life span is postreproduction. However, in daf-2 mutants, progeny are produced very late into the life span. This suggests that daf-2 animals do not live very long postreproduction; instead they show extended reproductive life spans albeit with lower fertility. They produce progeny up to the day they die. It is interesting to note that daf-2(e1391) animals exhibit a fourfold increase in life span, which is higher than reported for other daf-2 alleles (KENYON et al. 1993 Down; DORMAN et al. 1995 Down; LARSEN et al. 1995 Down; Table 4).

Longevity defects in daf-2 and age-1 are not due to reduced fertility:
We examined whether the increase in longevity in daf-2 mutants was associated with the reduced brood size of these mutants. Previously it had been reported that daf-2(e1370) had a wild-type brood size but a twofold increase in life span at 15° (LARSEN et al. 1995 Down). Therefore, we compared daf-2(e1370), daf-2 (e1391), and the backcrossed daf-2(e1370) mutants for total progeny produced as well as life span at 15° (Table 5). All three strains had nearly wild-type brood sizes (Table 5). However, several of the strains had increased life spans (Table 5). This suggests that the life span defect in daf-2 mutants is unrelated to the number of eggs produced, similar to the results with daf-2(e1370) of LARSEN et al. 1995 Down.

Because we observed differences in brood size when animals were shifted to the nonpermissive temperature at different stages of development, we examined whether life span was also affected in a similar manner. In Figure 3, we plot the life spans of wild-type N2, daf-2 (e1370), daf-2(e1391), and daf-2(e1370) backcrossed when shifted as L4 and young adult. For all strains there was no significant difference in life span between the L4 and adult temperature-shifted animals, suggesting again that there is no correlation between the number of progeny produced and life span.



View larger version (21K):
In this window
In a new window
Download PPT slide
  		Figure 3. Time of upshift does not affect longevity of temperature sensitive daf-2 mutants. Animals were grown at 15° until either the L4 stage or young adult (YA), when they were transferred to 25°, except for N2 which was grown at 20° and then shifted to 25°. Mean life spans for each strain are shown in Table 4. For each strain, multiple life span assays were done. The total number of animals counted in each assay are the following: for N2 L4 shift, n = 170; for N2 young adult shift, n = 36; for daf-2(e1370) L4 shift, n = 18, with an additional 34 animals (65%) that died from internal hatching of the progeny and young adult shift, n = 34, with an additional 39 animals (53%) that died from internal hatching of the progeny; for daf-2(e1391) L4 shift, n = 42, with an additional 29 animals (41%) that died from internal hatching of the progeny and young adult shift, n = 34, with an additional 29 animals (46%) that died from internal hatching of the progeny; for daf-2(e1370) backcrossed strain, n = 22, with an additional 33 animals (60%) that died from internal hatching of the progeny and young adult shift n = 21, with an additional 21 animals (50%) that died from internal hatching of the progeny. Animals continued to produce progeny late into the life span. daf-2(e1391) continued to produce progeny up to Day 46 of the life span. Life span assays were performed as outlined in MATERIALS AND METHODS.

daf-2 reduced brood size and lethality/arrest is suppressed by daf-16:
Because daf-16 suppresses all known daf-2 and age-1 phenotypes (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; KENYON et al. 1993 Down; GOTTLIEB and RUVKUN 1994 Down; DORMAN et al. 1995 Down), we examined whether a daf-16 mutation could restore wild-type fertility to daf-2 and age-1 mutants. The brood size of the daf-16(m27); daf-2(e1370) double mutant is significantly increased to a mean brood size of 189 (Table 6). daf-16(m27) mutants have a mean brood size of 202, which is slightly decreased from wild-type N2 of 256 (Table 6). daf-16(m27) only partially suppresses the dauer arrest phenotype of daf-2(e1391) (Table 6). The daf-16(m27); daf-2(e1391) double mutant has a mean brood size of 155, which is many more than daf-2 (e1391), but is still decreased compared to wild type. Thus, similar to all other phenotypes associated with daf-2 (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; GOTTLIEB and RUVKUN 1994 Down; DORMAN et al. 1995 Down), the fertility defect of daf-2 is at least partially suppressed by daf-16 (Table 6).


 
View this table:
In this window
In a new window

  		Table 6. daf-2 reduced brood size and lethality is suppressed by daf-16

The daf-2 larval arrest and embryonic lethal phenotype is completely suppressed by a mutation in daf-16. Even though daf-16(m27) is a partial suppressor of the dauer arrest and brood size defects of daf-2(e1391), it potently suppresses the arrest/lethality associated with daf-2(e1391) mutants at 25°, because none of the progeny laid arrested as L1/L2 or died as eggs in the double mutant (Table 6).

We further examined the ability of daf-16 to suppress the aging defects in both daf-2(e1370) and age-1(mg44) at 25°. L4 animals were shifted from 15° to 25° and life span was determined. The mean life span of both daf-16(m27); daf-2(e1370) and daf-16(m27); sqt-1 (sc13) age-1(mg44) were similar to wild type (data not shown). Therefore, daf-16 completely suppresses the life span defects in both age-1 and daf-2 mutants.

Both daf-2 and age-1 show a synthetic lethality with the mutations affecting the amphid processes (che/osm class of mutations; VOWELS and THOMAS 1992 Down; GOTTLIEB and RUVKUN 1994 Down). At the daf-2(e1370) nonpermissive temperature, the mean brood size of daf-2(e1370); osm-3(p802) mutants was 27, with 26% of the brood arresting as larvae or dying as eggs, when animals are shifted as young adults. Most of the arrested larvae could be recovered and form fertile adults if shifted to the permissive temperature, suggesting that the animals indeed are arrested and this is not a larval lethal interaction. The triple mutant, daf-16(m27); daf-2 (e1370); osm-3(p802) has a wild-type brood size with no associated arrest (data not shown). Therefore, daf-16 can suppress all the phenotypes of daf-2.

daf-16 is the major downstream target of daf-2:
In mam- malian cells, insulin receptor couples to many downstream targets that mediate several different signaling cascades, including PI-3-kinase, Grb-2, pp60 c-src, Shc and PLC gamma (reviewed in KAHN 1994 Down; KAHN and WEIR 1994 Down; BONFINI et al. 1996 Down). We wanted to determine if there were other negatively regulated downstream targets of the daf-2 insulin-like signaling pathway other than daf-16. Therefore, we screened for suppressors of the daf-2; daf-12 arrest phenotype.

Double mutants of both daf-2 and daf-12 at the nonpermissive temperature and age-1 and daf-12 result in an arrest at the L1 and L2 stage (VOWELS and THOMAS 1992 Down; GOTTLIEB and RUVKUN 1994 Down). daf-16 can completely suppress the growth arrest phenotype of both the daf-2; daf-12 and age-1; daf-12 double mutants (GOTTLIEB and RUVKUN 1994 Down). The daf-2(e1370); daf-12 (m20) double mutant strain provides for an easy selection for daf-2 or daf-12 suppressors. We screened 7000 genomes for suppressors of the daf-2; daf-12 arrest at 25° and found four independent alleles that allow reproductive development at 25°; mg51, mg52, mg53, and mg54. All four alleles map in the daf-16 genetic interval near dpy-5 on Chromosome I, and failed to complement daf-16, and were further characterized for dauer formation and suppression (GOTTLIEB and RUVKUN 1994 Down; OGG et al. 1997 Down; S. GOTTLIEB and G. RUVKUN, unpublished observations). This suggests that the major target for daf-2 is daf-16, because only alleles of daf-16 were identified. Other mutations that may suppress daf-2 are relatively rare or may not exist.


*  DISCUSSION
*TOP
 *ABSTRACT
 *MATERIALS AND METHODS
 *RESULTS
 *DISCUSSION
*LITERATURE CITED

An insulin-like signaling pathway controls C. elegans longevity, metabolism, and fertility:
Mutations in daf-2 or age-1 affect life span (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; KENYON et al. 1993 Down; DORMAN et al. 1995 Down; LARSEN et al. 1995 Down; MORRIS et al. 1996 Down; this study). We have shown that mutations in either of these two genes also affect fertility and embryonic development, and that all of these phenotypes are regulated via outputs to daf-16. The molecular identity of both daf-2 and age-1 as part of an insulin-like signaling pathway (MORRIS et al. 1996 Down; KIMURA et al. 1997 Down) suggests how these genes coordinately control longevity, metabolism, and fertility. The similar phenotypes and genetic epistasis results of daf-2 and age-1 position them at the same point in the dauer genetic pathway, consistent with their molecular identities in a similar signal transduction pathway (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; GOTTLIEB and RUVKUN 1994 Down; DORMAN et al. 1995 Down; LARSEN et al. 1995 Down; MORRIS et al. 1996 Down; KIMURA et al. 1997 Down). Therefore, we suggest that dauer diapause, longevity, reproduction, and embryonic development are regulated by an insulin-like signaling cascade through the DAF-2 insulin receptor family member and the AGE-1 PI-3-kinase.

Animals that arrest at the dauer diapause stage, such as daf-2 and age-1 mutants, show a decrease in the rate of metabolism, suspend development of both somatic and germline tissues, and age more slowly, allowing the animals to survive long periods under suboptimal conditions (reviewed in RIDDLE 1988 Down; THOMAS 1993 Down; RIDDLE and ALBERT 1997 Down). In this study, we have shown that daf-2 and age-1 animals have a decline in fertility when shifted to high temperatures even as young adults. It is possible that decreased insulin signaling in adult daf-2 animals causes metabolic changes as seen by the change in fat in daf-2 mutants shifted to a non-permissive temperature at this stage (KIMURA et al. 1997 Down). Indeed, both age-1 and daf-2 mutants show a metabolic shift as adults as examined by changes in the levels of many metabolic enzymes (VANFLETEREN and DE VREESE 1995 Down). Alternatively, the L4 or young adult temperature-shift experiment may uncover a defect, for example, in gonadogenesis, that may have occurred much earlier.

Mutations in daf-2 affect embryogenesis, reproduction, longevity, and dauer arrest. We have shown that several daf-2 alleles cause a small percentage of the animals to arrest as larvae or die as eggs. None of the daf-2 alleles generated so far are null alleles (KIMURA et al. 1997 Down). Therefore, it is possible that the null phenotype of daf-2 is more severe than nonconditional dauer arrest (i.e., embryonic or larval arrest). Indeed, null mutants of the mouse or human insulin receptor result in runaway lipolysis and associated ketoacidosis and neonatal death (KROOK et al. 1993 Down; WERTHEIMER et al. 1993 Down; ACCILI et al. 1996 Down; JOSHI et al. 1996 Down). The developmental arrest caused by vertebrate insulin receptor mutations may be due to the metabolic defects caused by lack of insulin signaling, or the arrest may be caused more directly by the lack of insulin growth factor responses. We suggest that daf-2 and age-1 also regulate growth of the germ line and embryo, as well as metabolism.

All of the phenotypes caused by mutations in either age-1 or daf-2, including the embryonic lethality and fertility defects, are suppressed by mutations in daf-16 (RIDDLE et al. 1981 Down; VOWELS and THOMAS 1992 Down; KENYON et al. 1993 Down; GOTTLIEB and RUVKUN 1994 Down; LARSEN et al. 1995 Down; DORMAN et al. 1995 Down; this study); daf-16 is the major downstream target of insulin-like signaling in C. elegans. daf-16 encodes a forkhead type transcription factor (OGG et al. 1997 Down). Therefore, insulin-like signaling pathways couple to the transcriptional activity of DAF-16 to in turn control fertility and metabolism. Because all of the daf-2 and age-1 mutant phenotypes are suppressed by daf-16 mutations, the other possible signaling outputs from the daf-2 insulin-like signaling cascade, for example, to glucose transport or metabolic enzyme activities, are not as germane to longevity and fertility as the daf-16 transcriptional output.

Our results do not support a model that fertility decline is the cause of longevity increase. We find that life span in these mutants can be decoupled from their effects on brood size, similar to previous reports of daf-2 (e1370) by LARSEN et al. 1995 Down. Therefore, the increased longevity seen in daf-2 and age-1 mutants is not due to the suspension of fertility. This is in contrast to studies in Drosophila where high rates of egg production decreased life span in females (PARTRIDGE et al. 1987 Down). Moreover, in female flies, life span was increased by exposure to higher temperature, and this correlated with a decrease in egg production at these higher temperatures (SMITH 1958 Down). Further studies revealed that life span of females can be increased by reducing the rate of egg-laying (SMITH 1958 Down).

In C. elegans, two studies that examine the relationship between mating and life span find that mating can reduce male life span (VAN VOORHIES 1992 Down). Mating to hermaphrodites either reduces (GEMS and RIDDLE 1996 Down) or has no effect (VAN VOORHIES 1992 Down) on life span. However, this reduction in life span is not due to an effect on either egg or sperm production, but rather a cost of copulation (GEMS and RIDDLE 1996 Down). In Drosophila, female life span is also decreased by courting (CHAPMAN 1992 Down) and receipt of male accessory fluid (CHAPMAN et al. 1995 Down); male life span is shortened by sexual activity (PARTRIDGE and FARQUHAR 1981 Down).

In wild-type animals, half of the life span is postreproductive, which is significantly different from daf-2 mutants where progeny can be produced up until the day of death. Brood size is limited by the number of sperm in a wild-type hermaphrodite (HODGKIN and BARNES 1992 Down). Adult hermaphrodites will lay unfertilized eggs at the end of the brood, and in fact brood size is significantly increased when hermaphrodites are mated to wild-type males (HODGKIN and BARNES 1992 Down). Because daf-2 mutants have such a small brood size, it is possible that they continue to have progeny up until they die because they do not run out of sperm. However, we have noted that even with the small brood size, daf-2 animals also lay unfertilized eggs (H. A. TISSENBAUM and G. RUVKUN, unpublished results). We hypothesize that the fertility defects in daf-2 and age-1 mutants are due to the metabolic shift towards energy storage so that fewer oocytes and sperm are generated.

age-1(hx546) affects maternal and zygotic age-1 activity:
age-1(hx546) animals are long-lived compared to wild type (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; DORMAN et al. 1995 Down; MALONE et al. 1996 Down; MORRIS et al. 1996 Down; this study). age-1(hx546) has reduced zygotic age-1 activity but age-1(hx546) is not dauer constitutive at most temperatures, because the zygotic age-1 gene activity of this allele is sufficient to allow nonarrested development (KLASS 1983 Down; FRIEDMAN and JOHNSON 1988A Down, FRIEDMAN and JOHNSON 1988B Down; MALONE et al. 1996 Down; MORRIS et al. 1996 Down; this study). However, the major defect in age-1 (hx546) is in maternal contribution of age-1 activity: at high temperature, age-1(hx546) shows no more age-1 activity than the age-1(mg44) molecular null allele (MORRIS et al. 1996 Down; this study). Thus the long life span of age-1(hx546) is caused by a defect in maternally contributed age-1 activity as well as a decline in zygotic age-1 activity.

The age-1 alleles (mg109, mg44, m333), including those that are predicted to truncate the AGE-1 protein, show a dauer-constitutive phenotype that can be rescued by wild-type maternal gene activity (GOTTLIEB and RUVKUN 1994 Down; MORRIS et al. 1996 Down). In addition, both age-1(m333) and age-1(mg44) allele, which are probable null alleles, show a dramatic