Identification of Ligands for DAF-12
that Govern Dauer Formation
and Reproduction in C. elegans
Daniel L. Motola,1Carolyn L. Cummins,1Veerle Rottiers,3Kamalesh K. Sharma,2Tingting Li,1Yong Li,4
Kelly Suino-Powell,4H. Eric Xu,4Richard J. Auchus,2Adam Antebi,3and David J. Mangelsdorf1,*
1Howard Hughes Medical Institute and Department of Pharmacology
2Department of Internal Medicine
University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75390, USA
3Huffington Center on Aging, Department of Molecular and Cellular Biology, Baylor College of Medicine,
Houston, TX 77030, USA
4Laboratory of Structural Sciences, Van Andel Research Institute, Grand Rapids, MI 49503, USA
In response to environmental and dietary cues,
the C. elegans orphan nuclear receptor, DAF-
12, regulates dauer diapause, reproductive de-
velopment, fat metabolism, and life span. De-
spite strong evidence for hormonal control,
the identification of the DAF-12 ligand has re-
mained elusive. In this work, we identified two
distinct 3-keto-cholestenoic acid metabolites
of DAF-9, a cytochrome P450 involved in hor-
mone production, that function as ligands for
DAF-12. At nanomolar concentrations, these
steroidal ligands (called dafachronic acids)
bind and transactivate DAF-12 and rescue the
hormone deficiency of daf-9 mutants. Interest-
ingly, DAF-9 has a biochemical activity similar
to mammalian CYP27A1 catalyzing addition of
olites. Together, these results define the first
steroid hormones in nematodes as ligands for
an invertebrate orphan nuclear receptor and
demonstrate that steroidal regulation of repro-
duction, from biology to molecular mechanism,
is conserved from worms to humans.
The genome of C. elegans is predicted to contain 284 nu-
clear receptors (Gissendanner et al., 2004; Sluder and
Maina, 2001). Forward and reverse genetic studies have
uncovered roles for these receptors in diverse physiolog-
ical processes such as development, reproduction, and
metabolism. However, all nuclear receptors in worms re-
main orphans, since ligands regulating their function
have not been identified (Sluder and Maina, 2001; Van
Gilst et al., 2005a; Van Gilst et al., 2005b).
In contrast to other C. elegans nuclear receptors, a con-
siderable amount of genetic evidence supports the exis-
tence of a steroid-like ligand for the orphan receptor,
DAF-12. daf-12 belongs to a group of over 30 genes, col-
lectively called daf (dauer formation) genes, which trans-
duce environmental signals that influence the choice be-
tween alternative developmental programs of dauer
diapause or reproductive development (Antebi et al.,
2000; Riddle and Albert, 1997). Dauer diapause occurs
when second-larval-stage animals (L2) delay further re-
productive development under conditions of diminishing
food or overcrowding and instead form the nonfeeding,
nonreproductive, and long-lived dauer larva (Riddle and
Albert, 1997). Under more favorable conditions, dauer lar-
Daf genes generally produce a dauer-constitutive pheno-
type (Daf-c) or a dauer-defective phenotype (Daf-d). Daf-c
mutants always arrest as dauers, while Daf-d mutants by-
pass dauer, regardless of environmental signals. Loss of
daf-12 results in Daf-d as well as L3 stage heterochronic
phenotypes, indicating that daf-12 is required for dauer
formation and for proper developmental timing in the re-
productive state (Antebi et al., 1998; Antebi et al., 2000).
Analysis of Daf genes has revealed that favorable envi-
ronments activate insulin/IGF-1 and TGFb signaling path-
ways within the organism that converge on DAF-12 to
inhibit its dauer-promoting function and activate its repro-
ductive function (Kimura et al., 1997; Ren et al., 1996;
Schackwitz et al., 1996). Acting cell nonautonomously,
these pathways are believed to directly or indirectly stim-
P450, DAF-9 (Gerisch et al., 2001; Jia et al., 2002). Evi-
dence for this model stems from the findings that insu-
lin-like receptor (daf-2), TGFb (daf-7), and cytochrome
of daf-12 (Vowels and Thomas, 1992; Thomas et al., 1993;
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1209
Gerisch et al., 2001). In addition, Daf-c mutants of daf-12
have been isolated that map to a single residue (R564) in
the putative ligand binding domain of DAF-12 and are pre-
dicted to perturb ligand binding (Antebi et al., 2000). Phe-
notypically, these mutants arrest as partial dauers but
recover and resemble weak daf-9 alleles that exhibit go-
nadal migration (Mig) defects (Gerisch et al., 2001; Jia
et al., 2002). Thus, the predicted loss of hormone produc-
tion in daf-9 null worms or loss of hormone binding by
daf-12 Daf-c worms results in the failure to inhibit dauer-
promoting functions and activate L3 stage reproductive
functions of DAF-12.
Several lines of evidence suggest that DAF-12 ligands
may be derived fromcholesterol and function aseither en-
docrine or paracrine hormones. First, C. elegans lacks the
ability to synthesize cholesterol (Chitwood, 1999), and
cholesterol deprivation produces Mig and Daf-c pheno-
types (Gerisch et al., 2001; Jia et al., 2002; Matyash et al.
2004). Second, worms missing both homologs (ncr-1,
ncr-2) of the human Niemann-Pick type C1 gene, a mem-
brane glycoprotein implicated in cholesterol transport, ar-
rest constitutively as dauers (Li et al., 2004). Furthermore,
recent evidence shows lipid extracts from wild-type
worms can rescue daf-9 phenotypes (Gill et al., 2004). Fi-
nally, unlike DAF-12, DAF-9 functions non-cell autono-
mously and is restricted to a few cell types such as XXX
cells, which are believed to be neuroendocrine cells and
a main source of DAF-12 ligands (Ohkura et al., 2003). Al-
derived hormone promotes reproductive development in
C. elegans, the identities of DAF-9-derived hormonal li-
gands that activate DAF-12 have remained elusive.
In this study, we identified a group of 3-keto-sterols that
serve as DAF-9 substrates and are metabolized through
gands rescue the Daf-c phenotypes displayed by muta-
tions in daf-9, daf-2, daf-7, and ncr-1;ncr-2. Finally, frac-
tionation of C. elegans lipid extracts shows that these
metabolites exist in vivo but are absent in daf-9 null
worms. These data provide unequivocal evidence for the
existence of nematode hormones and demonstrate the
evolutionary conservation of broad aspects of endocrine
steroid signaling from worms to human.
Identification of 3-Keto-Steroids
as DAF-12 Activators
DAF-12 ligand screening was performed using a chimeric
GAL4-DAF-12 cotransfection assay in HEK293 cells. As-
says were performed in the presence or absence of co-
and Antebi, 2004; Gerisch et al., 2001; Gill et al., 2004; Jia
et al., 2002; Mak and Ruvkin, 2004). Our initial compound
screen included bile acids, steroids, and lipophilic com-
pounds that are known ligands for PXR, VDR, and LXR,
the closest vertebrate homologs of DAF-12 (Antebi et al.,
2000; Mooijaart et al., 2005). This screen identified
3-keto-lithocholic acid (3K-LCA, Figure 1A) as a weak
activator of DAF-12 independent of DAF-9 (Figure 1B).
Lithocholic acid (LCA), which differs from 3K-LCA by an
a-hydroxyl group at C-3, did not exhibit activity on its own
or in the presence of cotransfected DAF-9 (Figure 1B).
These results suggested that a C-3 ketone was required
for DAF-12 activation by 3K-LCA. No other bile acids or
steroids tested activated DAF-12, including cholic acid
(CA), chenodeoxycholic acid (CDCA), deoxycholic acid
estradiol, corticosterone, 1,25-dihydroxyvitamin D3, and
20-hydroxyecdysone (Figure 1B and data not shown).
The above data suggested that endogenous 3-keto-
sterols from C. elegans are candidate DAF-12 ligands.
Lathosterol and its 4-methyl-derivative, lophenol, are cho-
lesterol metabolites that have distinct effects on the nem-
Merris et al., 2003). When given as the sole dietary sterol,
lathosterol supported full reproductive growth (Merris
etal.,2003),while worms grownonlyin thepresence of lo-
phenol constitutively entered dauer diapause (Matyash
et al., 2004). These studies suggest lathosterol but not lo-
phenol may be a direct precursor to the DAF-12 ligand.
Therefore, we tested lathosterol and lophenol along with
none, for activity in the cotransfection assay. Upon co-
transfection with DAF-9, activation of DAF-12 was mark-
edly increased by lathosterone (433-fold) and lophenone
(103-fold), but not by their respective 3b-hydroxy deriva-
tives (Figure 1C). In addition, 4-cholesten-3-one, a natural
oxidation product of cholesterol, activated DAF-12 (109-
fold) in the presence of DAF-9 (Figure 1C). Unlike 3K-LCA,
activation of DAF-12 by lathosterone, lophenone, and
turally, 3K-LCA differs from these 3-keto-sterols in the
length and oxidation state of the side chain and in the sat-
gest that DAF-9 converts 3-keto-sterols, possibly through
side chain oxidation, into DAF-12 activators.
DAF-9 Derivatives of 3-Keto-Sterols Rescue daf-9
To determine whether the 3-keto-sterol metabolites were
biologically relevant we tested their ability to rescue the
Daf-c and Mig phenotypes of daf-9 null animals (Albert
and Riddle, 1988; Gerisch et al., 2001; Jia et al., 2002).
Individual 3-keto sterols or their respective 3b-hydroxy
sterols were incubated with Sf9 cell microsomes contain-
ing DAF-9 and the human P450 oxidoreductase (hOR),
somes. As a control, microsomes from cells expressing
hOR alone were used. For rescue experiments we utilized
daf-9(dh6)dhEx24 worms, which carry an unstable extra-
chromosomal array of daf-9(+) linked to the nuclear
marker sur-5::gfp (Gerisch et al., 2001). The progeny of
these animals, which contain a mixture of both daf-9(+)
1210 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
Figure 1. Activation of DAF-12 by 3-Keto-Sterols Requires DAF-9
(A) Structures of candidate DAF-12 ligand precursors relative to cholesterol and 3-keto-lithocholic acid.
(B and C) Activation of DAF-12 by 10 mM bile acids (B) or C. elegans sterols (C) and their 3-keto derivatives in the presence of DAF-9 (black bars) or an
emptyCMX control (white bars) vector. In (B), cotransfection of the intestinal bile acid transporter (IBAT) expression plasmid wasused to facilitate bile
(D) Rescue of daf-9(dh6) null worms by sterols after incubation with DAF-9 microsomes. Results are reported as percentage of animals rescued from
dauer as wild-type gravid adults (wt), Mig adults, or molt-defective larvae. Numbers in each bar refer to worms tested.
(E) Dose response of DAF-12 activation to indicated sterols in cells cotransfected with DAF-9.
Abbreviations: chenodeoxycholic acid (CDCA), cholic acid (CA), deoxycholic (DCA) acid, lithocholic acid (LCA), 3-keto-lithocholic acid (3K-LCA),
6-keto-lithocholic acid (6K-LCA), 7-keto-lithocholic acid (7K-LCA). Reporter gene activity is expressed as fold induction of relative light units
(RLU) compared to ethanol control (n = 3 ± SD).
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1211
tracts for 48 hr. daf-9(?) worms were then isolated and
scored 24 hr later for presence of the dauer phenotype.
Extracts from DAF-9 microsomes incubated with either
4-cholesten-3-one or lathosterone resulted in 100% res-
cue of the Daf-c and Mig phenotypes in daf-9(?) animals
(Figure 1D). Remarkably, these animals were indistin-
guishable from wild-type adults. Indeed, they bypassed
dauer, exhibited normal gonadal migration, and produced
all Daf-c progeny upon passage to plates lacking micro-
somal extracts. Rescue by 4-cholesten-3-one and lathos-
terone required conversion by DAF-9, since their incuba-
tion with control microsomes did not rescue any daf-9
phenotypes (data not shown). Extracts from DAF-9 micro-
somes incubated with lathosterol (which does not con-
tain a 3-keto group) partially rescued (83%) the Daf-c
phenotype, but these animals were Mig and sterile (Fig-
ure 1D). This effect also required DAF-9 and was not
seen using control microsomes (data not shown). Finally,
resulted in rescue of Daf-c, none of these animals were
normal (58% were Mig; 42% failed to enter dauer, had
molting defects, or were dead; Figure 1D). This effect
was dependent on DAF-9 and the C-3 ketone in lophe-
none, since no effect was seen using lophenol or control
microsomes. Also, no effects were seen using cholesterol
as a substrate. Altogether, these assays revealed that
DAF-9 microsomes convert 4-cholesten-3-one and lath-
osterone into activities that completely rescued daf-9
Identification of DAF-9 Metabolites
The data above suggested that DAF-9 has enzymatic ac-
tivity that converts 4-cholesten-3-one and lathosterone
into DAF-12 ligands. Although the absolute potency of
these DAF-9 products could not be determined using the
rescue assay, dose-response curves from the cotransfec-
tion assay revealed that lathosterone metabolites were ei-
ther significantly more potent or more abundantly pro-
duced than 4-cholesten-3-one metabolites (Figure 1E).
To identify these DAF-9 products, we utilized liquid
chromatography/mass spectrometry (LC/MS). The 3-
keto-D4-enone structure present in 4-cholesten-3-one
has significant UV absorbance at 240 nm, permitting de-
tection of the metabolites. Incubation of 4-cholesten-3-
one with DAF-9 microsomes yielded two new peaks at
240 nm that were not present in the control microsomes
(Figure 2A). These products eluted at 4.0 (peak 1) and 5.5
(peak 2) minutes on a reverse phase C18column, raising
lesten-3-one, which elutes at 12.5 min (Figure 2A). Since
lathosterone is not UV active, its DAF-9 metabolites were
scanned in negative-ion mode, revealing two peaks that
were not present in the control microsome reactions.
DAF-9 products of lathosterone eluted much earlier than
lathosterone at4.0 and 4.2 min (peaks 3and 4),analogous
to the pattern seen for 4-cholesten-3-one (Figure 2B and
data not shown). Fractions from DAF-9 microsomal reac-
tions subjected to reverse-phase HPLC were also tested
ures 2C–2F). Fractions corresponding to peaks 1–4 (i.e.,
the 4-cholesten-3-one and lathosterone derivatives) acti-
vated DAF-12 several hundred-fold in the absence of co-
transfected DAF-9 (Figures 2C and 2D) and rescued daf-
9 null animals (Figures 2E and 2F). Again, the lathosterone
derivatives were stronger at activating DAF-12 and rescu-
ing daf-9, suggesting these compounds may be more
abundant, potent, or efficacious.
Next, we utilized LC/MS as a first step in the identifica-
tion of the activities in peaks 1–4. Based on the molecular
weight of 4-cholesten-3-one ([M + H]+m/z = 385) and the
retention times and mass spectra of the new compounds,
peak 2 is consistent with a monohydroxylated derivative
of 4-cholesten-3-one ([M + H]+m/z = 401) and peak 1 is
a carboxylic acid derivative was found after a negative-ion
scan in which a unique peak at 4.0 min was found only
in the DAF-9 microsomes with a base peak at m/z 413
(Figure 2G). Peak 4, which was scanned in negative ion
mode, yielded similar mass spectra to peak 1, consistent
with the conclusion that peak 4 is the carboxylic acid
derivative of lathosterone (Figure 2H). Finally, peak 3 con-
tained one DAF-9-specific product at m/z 415 (Figure 2H).
Although the identity of this peak remains unknown, the
activity we observed in fractions 4 and 5 (from Figures
2B and 2D) tracks predominantly with peak 4 and not
peak 3 (see Figure S1 in the Supplemental Data available
with this article online).
DAF-9 Is a 3-Keto-Sterol-26-Monooxygenase
The finding that 3K-LCA was a weak activator of DAF-12
suggested that the position of oxidation of DAF-9 metab-
olites was on the sidechain. Thecommercial availability of
monohydroxylated derivatives of cholesterol permitted us
to focus on defining the site of oxidation on the 4-choles-
ten-3-one metabolites of DAF-9. A panel of side chain ox-
idized 4-cholesten-3-one derivatives was generated by
converting 5-cholesten-3b-ol (ring A, 3b-hydroxy-D5) oxy-
sterols (20(S)-OH-, 22(R)-OH-, 22(S)-OH-, 24-OH-, 25-
OH-, (25R),26-OH-, and (25S),26-OH-cholesterol) into
their respective 4-cholesten-3-one (ring B, 3-keto-D4)
oxysterols using cholesterol oxidase (Figure 3A). When
tested in the cotransfection assay without DAF-9, two di-
astereomers of 26-hydroxy-4-cholesten-3-one ((25S),26-
hydroxy-4-cholesten-3-one and (25R),26-hydroxy-4-cho-
lesten-3-one) were strongly active (Figure 3B). In contrast,
the ring A oxysterols were inactive, confirming the idea
that DAF-12 ligands require a 3-keto group. Chromato-
graphic separation of the ring B oxysterols and compari-
son to the 4-cholesten-3-one metabolites of DAF-9 re-
solved peak 2 (Figure 2A) into two peaks that coeluted
with the diastereomers of 26-hydroxy-4-cholesten-3-one
(Figure 3C). Theseresults revealed that DAF-9 isa nonste-
reoselective 4-cholesten-3-one 26-hydroxylase.
1212 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
Figure 2. Identification of Carboxylated Metabolites of 4-Cholesten-3-One and Lathosterone as DAF-12 Agonists
(A) Representative UV chromatogram of 4-cholesten-3-one and (B) reconstructed total-ion-current chromatogram of lathosterone after incubation of
100 mM substrate with DAF-9 (black line) or control (red line) microsomes. Product peaks unique to DAF-9 and their retention times are shown in blue.
IS, internal standard of 1,4-cholestadien-3-one.
B). Transfections and rescue assays are described in Figure 1 legend. Average number of worms tested in (E) and (F) were 75 and 125, respectively.
(G and H) Mass spectra of DAF-9 metabolites of 4-cholesten-3-one (peaks 1 and 2) and lathosterone (peaks 3 and 4) scanned from m/z 250–500.
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1213
Given that DAF-9 microsomes produced both hydroxyl-
ated and carboxylated metabolites of 4-cholesten-3-one,
we investigated whether DAF-9 oxidizes 4-cholesten-3-one
at C-26 to produce both diastereomers of 26-hydroxy-4-
Indeed, incubation of either (25S) or (25R),26-hydroxy-
4-cholesten-3-one with DAF-9 microsomes resulted in
the production of a single UV active peak at 4 min and
Figure 3. DAF-9 Is a 3-Keto-Sterol C-26 Monooxygenase
(A) Side chain substitutions of 5-cholesten-3b-ol (ring A) and 4-cholesten-3-one (ring B) derivatives.
(B) DAF-9-independent activation of GAL4-DAF-12 in HEK293 cells after incubation with the indicated sterols (10 mM for all sterols except 22(R)-hy-
droxy-4-cholesten-3-one, which was 4 mM). n = 3 ± SD.
(C) UV chromatogram of 4-cholesten-3-one oxysterols (top panel) compared to DAF-9 (black line) or control (red line) microsomes incubated with
(D) UV chromatogram of DAF-9 (black line) and control (red line) microsomes after incubation with 100 mM (25R),26-hydroxy-4-cholesten-3-one (top
panel) or (25S),26-hydroxy-4-cholesten-3-one (bottom panel). Arrows indicate products unique to DAF-9 microsomes. Mass spectra of DAF-9 me-
tabolites (insets) were obtained in positive ion scan mode.
1214 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
m/z 415 in positive-ion mode (Figure 3D). This retention
time and mass spectral property were identical to the car-
boxylated metabolite found after incubation 4-cholesten-
3-one with DAF-9 microsomes (Figure 2A, peak 1) and
were not detected in control microsomal reactions (Fig-
ure 3D). In this context, DAF-9 substrate specificity re-
quired a 3-keto-D4structure, since the 3b-hydroxy-D5-
sterols, (25R) and (25S),26-hydroxycholesterol were not
oxidized (data not shown). Finally, incubation of either
sterol (25R),26-hydroxy-4-cholesten-3-one (R-OH) or
(25S),26-hydroxy-4-cholesten-3-one (S-OH) with DAF-9
microsomes resulted in complete rescue of Daf-c and
Mig phenotypes in 100% of animals tested (Figure 3E). In
contrast, extracts from control microsomes incubated
with the (25R)-sterol had no effect, while the (25S)-sterol
caused an incomplete rescue (10% molt defects, 75%
sterile Mig adults) resembling the activity in peak 2 of
Figure 2A (Figure 3E). These results demonstrated that
tivities through successive oxidations at C-26, resulting in
the production of carboxylic acid metabolites.
DAF-9 and Mammalian CYP27A1 Are Functional
Like DAF-9, successive oxidation of sterol substrates at
C-26 has been demonstrated for the mammalian cyto-
chrome P450, CYP27A1 (Cali and Russell, 1991). Notably,
ten-3-one more efficiently than cholesterol (Norlin et al.,
2003). Therefore, we tested the ability of CYP27A1 to
transfection of HEK293 cells with GAL4-DAF-12, human or
activation of GAL4-DAF-12 (Figure S2). In contrast, DAF-
of bovine adrenodoxin alone. Interestingly, cotransfection
of CYP27A1 in the presence of 25mM lathosterone (which
is a very good DAF-9 substrate) had no effect. Therefore,
DAF-9 and CYP27A1 have similar enzymatic activities
and overlapping, but distinct, substrate specificities.
3-Keto-4-Cholestenoic Acid Is a Hormonal Activator
To confirm C-26 carboxylic acids of 4-cholesten-3-one
(i.e., 3-keto-4-cholestenoic acid) as the more potent
DAF-9 metabolites, we synthesized their diastereomers
(Figure 4A; Table S1) and tested their ability to transacti-
vate DAF-12. The synthetic compounds exhibited chro-
matographic and mass spectral properties identical to the
acidicmetabolites obtained fromDAF-9 microsomes (Fig-
ure S3). When tested in the cotransfection assay, DAF-12
responded to all four steroids with the following rank order
of potencies: (25S),26-3-keto-4-cholestenoic acid (EC50=
100 nM); (25R),26-3-keto-4-cholestenoic acid (EC50R 1
mM); (25S),26-hydroxy-4-cholesten-3-one (EC50R 1 mM);
(25R),26-hydroxy-4-cholesten-3-one (EC50R 2 mM) (Fig-
ure 4B). Activation by (25S),26-3-keto-4-cholestenoic
acid was specific to DAF-12 and not observed with other
C. elegans, Drosophila, and human nuclear receptors (Fig-
ure S4). In addition, the DAF-12 ligand binding domain
mutants (R564C, R564H) were dramatically attenuated
in their response to (25S),26-3-keto-4-cholestenoic acid
(Figure S4) and did not respond to (25R),26-3-keto-4-cho-
lestenoic acid (data not shown).
Consistent with its function as a DAF-12 hormonal li-
gand, (25S),26-3-keto-4-cholestenoic acid rescued Daf-
c and Mig daf-9 phenotypes (Figures 4C and 4D). At
hormone concentrations of 250 nM, daf-9 animals were
indistinguishable from wild-type: they bypassed dauer
diapause to become reproductive adults (Figures 4C
and 4D) and reproduced like wild-type (?300 offspring;
n = 5 worms). Rescued animals also had normal gonads,
lacked dauer alae, and displayed normal pharyngeal
expansion (Figure 4C). Without hormone, these animals
produced all dauer progeny, confirming their daf-9 geno-
type and demonstrating a lack of maternal rescue (data
daf-9(rh50) was reversed at 250 nM hormone (Figure 4E).
At intermediate concentrations (50–100 nM), a proportion
of null mutants exhibited Mig and molting defects, sug-
gesting these phenotypes arise from a reduction in hor-
mone levels (Figure 4D and data not shown). The 25R di-
astereomer of 3-keto-4-cholestenoic acid also rescued
daf-9 phenotypes, albeit at5- to 10-fold higher concentra-
tions (Figure 4D). Consistent with the data above
(Figure 3E), (25R),26-hydroxy-4-cholesten-3-one had no
effect on daf-9 nulls, even at concentrations as high as
93% of worms (n> 60)to bypass dauer but still exhibit Mig
and/or molting defects and sterility (data not shown).
As predicted by cotransfection assays (Figure S4), daf-
12 LBD mutants were compromised in their ability to re-
spond to (25S),26-3-keto-4-cholestenoic acid (Figure 4E).
Hormone had no effect on daf-12(rh61), which truncates
types (Figure 4E). Another mutant, daf-12(rh273) contains
a missense lesion at a predicted ligand contact site and
gives Daf-c, Mig, and molting defects. Interestingly, at R
250 nM (25S),26-3-keto-4-cholestenoic acid, Daf-c but
not Mig or molt phenotypes were rescued (Figures 4E
and 4F and data not shown), implying this mutation de-
creases ligand binding affinity. Together, these results re-
vealed that 3-keto-4-cholestenoic acid can function as
a C. elegans hormone that inhibits dauer formation and
promotes reproductive development.
3-Keto-4-Cholestenoic Acid Acts Downstream
of Insulin, TGFb, and Cholesterol Lysosomal
To further investigate the biological activity of 3-keto-4-
cholestenoic acid, we tested its ability to rescue the dauer
constitutive phenotypes of worms carrying mutations
in the insulin-like receptor (daf-2), TGFb (daf-7), and the
Niemann-Pick type C1 homologs (ncr-1;ncr-2) positioned
upstream of daf-9 and daf-12. Accordingly, (25S),26-3-
keto-4-cholestenoic acid completely rescued the Daf-c
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1215
phenotypes of all these mutants (Figure 4F). However,
the stronger daf-2 mutant (e1370) circumvented dauer
morphogenesis but remained developmentally arrested
as dark L3-like larvae, consistent with the similar pheno-
type seen in daf-2;daf-12 double mutants (Vowels and
ing must impinge upon the pathway both upstream,
downstream, or parallel to hormone production.
Figure 4. 3-Keto-4-Cholestenoic Acid Is a Hormonal Ligand of DAF-12
(A) Structures of 4-cholesten-3-one metabolites of DAF-9
(B) Dose response of GAL4-DAF-12 activation to 4-cholesten-3-one metabolites in HEK293 cells (n = 3 ± SD).
(C) DIC microscopy of daf-9(dh6) (a-f) and daf-9(rh50) (g and h) mutants treated with or without 250 nM (25S),26-3-keto-4-cholestenoic acid. (a) Res-
cued adult,(b)partialdauer,(c)headofrescuedL3larva,(d)headofpartial dauer,(e)cuticleofrescued L3larva,(f)daueralae, (g)reflexedgonadofL3
larva, (h) unreflexed gonad of L3 larva.
(D) Response of daf-9(dh6) nulls treated with (25S),26-3-keto-4-cholestenoic acid or (25R),26-3-keto-4-cholestenoic acid. Results expressed as per-
centage of worms rescued from dauer after 3 days at 20ºC. Worms were scored as adults or molt-defective larvae.
(E) Rescue of Mig phenotypes by (25S),26-3-keto-4-cholestenoic acid. Results expressed as percentage of reflexed gonadal arms scored after
3 days at 20ºC (n > 60 from at least two independent experiments ± SD).
(F) Rescue of dauer phenotypes by (25S),26-3-keto-4-cholestenoic acid. Also shown by different shading are the percentage of dauer-rescued
worms that exhibited wild-type adult (black bar) or Mig (striped bar) gonads or an arrested L3 phenotype (white bar). Dauer rescue was scored after
2 days at 25ºC (daf-2 and daf-7) or 3 days at 20ºC (daf-9, daf-12, and ncr-1;ncr-2). n > 200 from at least two independent experiments ± SD.
1216 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
3-Keto-Cholestenoic Acids Bind DAF-12 as Bona
A hallmark of nuclear receptor agonists is their ability to
diametrically regulate interactions with corepressors and
coactivators. To test this, we utilized a mammalian
GAL4-coregulator/VP16-receptor two-hybrid assay in
HEK293 cells. In the absence of ligand, DIN-1S, a putative
DAF-12 corepressor (Ludewig et al., 2004), interacted
with DAF-12 as predicted (Figure 5A). Addition of 1 mM
(25S),26-3-keto-4-cholestenoic acid completely abol-
ished this interaction, supporting the conclusion that the
hormone disrupts dauer-promoting complexes involving
DAF-12 and DIN-1S. Hormone did not disrupt the interac-
tion between DIN-1S and a DAF-12-R564C mutant. We
also observed that DIN-1S-dependent repression of
DAF-12 basal activity could be reversed by addition of
100 nM (25S),26-3-keto-4-cholestenoic acid (Figure 5B).
In a similar coactivator interaction assay, (25S),26-3-keto-
4-cholestenoic acid induced the interaction between the
fourth receptor interaction domain (ID4) of the mammalian
coactivator protein SRC-1 and DAF-12, but not mutant
DAF-12-R564C (Figure 5C). We also tested transacti-
vation of full-length DAF-12 on a luciferase reporter plas-
mid containing the DAF-12 binding sites of lit-1 kinase,
Figure 5. 25(S),26-3-Keto-4-Cholestenoic Acid Functions as a Classical Nuclear-Receptor Ligand
(A) Ligand-dependent interaction of DAF12 with DIN-1S by mammalian two-hybrid analysis.
(B) Effect of DIN-1S on DAF-12 basal activation with (+) or without (?) 100 nM ligand. Cells were transfected with 45 ng/well DIN-1S and 15 ng/well
(C) Ligand-dependent interaction of DAF12 with SRC-1(ID4) by mammalian two-hybrid analysis.
(D) Ligand-dependent activation of full-length DAF-12 on a lit-1 kinase reporter gene.
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1217
a proposed DAF-12 target gene (Shostak et al., 2004). In
the presence of (25S),26-3-keto-4-cholestenoic acid,
DAF-12 transactivated the lit-1 kinase reporter plasmid in
a dose-dependent manner (EC50= 100 nM; Figure 5D).
tects ligand-dependent interactions between receptors
and coactivator peptides (Xu et al., 2002). At 1 mM, the
58-fold and 24-fold increases in binding units, respec-
tively, compared to vehicle control(Figure 6A).Incontrast,
the precursors, 4-cholesten-3-one and (25S),26-hydroxy-
4-cholesten-3-one, showed no significant binding. Weak
ing kinetics revealed that (25S),26-3-keto-4-cholestenoic
acid binds DAF-12 with high affinity (EC50 = 1 nM;
Figure 6B). A similar analysis showed (25R),26-3-keto-4-
finity (data not shown). Although the synthetic version of
the lathosterone carboxylic acid (3-keto-7,(5a)-choleste-
noic acid) is not available, strong ligand binding activity
was detected in DAF-9 microsomes that were incubated
with lathosterone (Figure 6C) and contain the carboxylic
that both the 3-keto-4-cholestenoic acid and 3-keto-7,
(5a)-cholestenoic acid hormones (Figure 6D)mediate their
effects in vivo through direct binding to DAF-12.
3-Keto-Cholestenoic Acids Are Endogenous,
A key prediction to the hypothesis that the 3-keto-choles-
tenoic acids are endogenous DAF-12 ligands is that they
should be present in wild-type but not daf-9 null worms.
To that end, crude lipid extracts from both wild-type and
daf-9 null (daf-9(e1406);daf-12(m20)) animals were gener-
ated asdescribedin theExperimental Procedures and an-
alyzed for GAL4-DAF-12 activity in HEK293 cells. Wild-
type worm extracts had strong DAF-12 activity, while as
expected no activity was detected from daf-9 null animals
(Figure S5). Wild-type and daf-9 extracts were then dis-
solved in methanol, fractionated by reverse-phase
HPLC, and tested again for activity. DAF-12 activity was
found only in HPLC fractions 4 and 5 from wild-type lipids
but not daf-9 null lipids (Figure 7A). Analysis of fractions 4
and 5 by selective ion monitoring (SIM) LC/MS showed
that the activity was specifically associated with a peak
at m/z 413 that was not detected in the inactive fractions
from daf-9 null animals (Figure 7B). The peak at m/z 413
was identical in mass and HPLC retention time to
3-keto-7,(5a)-cholestenoic acid (thecarboxylic acid deriv-
ative of lathosterone) that was produced by DAF-9
(Figure 2B). The estimated endogenous concentration of
this activity is ?200 nM (see Experimental Procedures
for details), which is well within the predicted limit for acti-
vation of DAF-12. Although further purification was re-
quired to detect the 3-keto-4-cholestenoic acid (see
Figure 6. DAF-9 Metabolites of 4-Cholesten-3-One and Lathosterone Bind DAF-12 as High Affinity Ligands
(A–C) AlphaScreen assay for ligand-dependent coactivator recruitment to the DAF-12 ligand binding domain. Reactions were performed in the pres-
ence of the indicated sterols (1 mM) (A), increasing concentrations of (25S),26-3-keto-4-cholestenoic acid (B), or a 1:5000 dilution of DAF-9 or control
microsomes incubated with 100 mM lathosterone (C). Results expressed as arbitrary binding units from triplicate assays (±SD).
(D) Structures of DAF-12 ligands.
1218 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
below), the fact that daf-9 null animals have no detectable
DAF-12 activity in any of the HPLC fractions indicates
daf-9 null worms lack significant amounts of either of the
DAF-12 ligands, including 3-keto-4-cholestenoic acid.
To provide further evidence for the presence of the 3-
tracts from L3- to L4-staged worms. Crude extracts were
fractionated bysilica column chromatography (Figure 7C),
and the DAF-12 activating fraction was determined to be
in the acetone:methanol eluate (Figure 7D). Subsequent
fractionation of this activity revealed the presence of two
distinct DAF-12 activity peaks at fractions 30–33 and
57–63 (Figure 7E). Although the level of activity in fraction
30–33 was too low for further analysis, its chromato-
graphic properties were consistent with the alcohol deriv-
atives of 4-cholesten-3-one. However, after pooling and
repurification by HPLC of fractions 57–63 enough material
was obtained to identify the carboxylic acid derivatives by
LC/MS (Figure 7F, upper panel). SIM mode identified
a peak at m/z 413 in negative-ion mode with a retention
time similar to the 3-keto-4-cholestenoic acid metabolite
of DAF-9 (Figure 7F, middle panel). This signal correlated
with DAF-12 activity, as it was not present in neighboring
fractions that lacked activity (data not shown). As ex-
pected, a second peak of higher abundance was also de-
tected that comigrated with the lathosterone metabolite,
3-keto-7,(5a)-cholestenoic acid (Figure 7F, bottom panel).
Although the concentration of 3-keto-4-cholestenoic acid
could not be determined with accuracy, its relative abun-
dance in the LC/MS indicates the in vivo concentration is
less than the 3-keto-7,(5a)-cholestenoic acid. These
pooled fractions rescued the Daf-c and Mig phenotypes
in 100% of daf-9 null worms tested (n > 300). Taken to-
gether, these results provide further evidence that both
of the 3-keto-cholestenoic acids (Figure 6D) are endoge-
nous hormonal ligands of DAF-12.
Discovery of Ligands for a C. elegans Orphan
In this paper, we detail the identification of 3-keto-4-cho-
lestenoic acid and 3-keto-7,(5a)-cholestenoic acid as en-
are 3-keto, C-26 oxidized derivatives of cholesterol that
differ in the position of an unsaturated double bond at
ern dauer formation and heterochronic developmental
Figure 7. 3-Keto-Cholestenoic Acids Are Endogenous Hormones
(A) DAF-12 activation by HPLC fractions from crude lipid extracts from wild-type and daf-9(e1406);daf-12(m20) worms.
(B) LC/MS chromatograms of the HPLC fractions 4 and 5 in (A).
(C) Strategy used to purify endogenous DAF-12 agonists from C. elegans lipid extracts.
(D) DAF-12 activation by fractionated lipid extracts.
(E) DAF-12 activation by silica column fractions of lipids eluted with acetone:methanol.
(F) LC/MS analysis of pooled and repurified fractions 57–64 in negative SIM mode (m/z 413) compared with DAF-9 metabolites of 4-cholesten-3-one
(middle panel) and lathosterone (bottom panel).
RLU, relative light units; n = 3 ± SD.
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1219
pathways, we propose to name these hormones D4-da-
fachronic acid and D7-dafachronic acid (Figure 6D).
Consistent with other nuclear-receptor ligands, the ste-
reochemistry of the DAF-12 ligand is an important deter-
minant for binding. Thus, the (25S),26-carboxylic acid
is ?10-fold more potent than the (25R),26-carboxylic
acid as a DAF-12 ligand. Although most of the studies
presented here focused on D4-dafachronic acid (i.e., 3-
keto-4-cholestenoic acid), we provide strong evidence
to suggest that the C-26 carboxylic acid metabolite of
lathosterone (i.e., D7-dafachronic acid) is also a bona fide
DAF-12 ligand. Like D4-dafachronic acid, the D7-dafach-
ronic acid was shown to bind and transactivate DAF-12,
rescue daf-9 worms, and was present in the most active
DAF-9 metabolites and lipid extracts from wild-type (but
not daf-9 null) worms at physiologically relevant concen-
trations. In fact, the activity and concentration of D7-da-
fachronic acid suggestthatit ismoreabundant andeffica-
the potency of D7-dafachronic acid as a DAF-12 ligand
awaits its de novo chemical synthesis. Taken together,
the dafachronic acids identified here represent the first li-
gands for a nematode nuclear receptor and are the first
steroid hormones identified in C. elegans.
An interesting aspect of this work is the finding that mul-
tiple DAF-12 ligands may exist. The ligands identified by
our approach were shown to exist in the most active lipid
fractions, but the possibility remains that other chemically
similar ligands may exist in these fractions as well. Our
work provides evidence for at least two ligands, raising
the intriguing possibility that each hormone may govern
the dafachronic acids have similar activities in preventing
dauer formation and promoting gonadal migration. How-
ever, it remains possible that these or other ligands differ-
entially affect other DAF-12-dependent functions, such as
life span. For example, DAF-9 and DAF-12 have been
shown paradoxically to be required for both inhibition and
extension of life span (Gerisch et al., 2001; Jia et al., 2002;
Larsen et al., 1995; Hsin and Kenyon, 1999; Gems et al.,
1998). Future work is being directed toward determining
whether the two dafachronic acids may be responsible
for these opposing activities or whether other discriminat-
ing ligands also exist. In addition, these ligands should
provide useful tools for characterizing DAF-12 target
genes and the biology of this hormone-receptor system.
Evolutionary Conservation of Steroid Hormone
Since its discovery, DAF-9 has been postulated to play
a key role in the synthesis of a cholesterol-derived hor-
mone that serves as a DAF-12 ligand and controls dauer
formation and reproductive development in C. elegans in
an endocrine fashion (Gerisch et al., 2001; Jia et al.,
2002; Gerisch and Antebi, 2004; Mak and Ruvkin, 2004).
The workpresented here provides thefirst direct evidence
for the involvement of DAF-9 in a steroidogenic pathway
that is conserved from worms to mammals. As with mam-
malian steroidogenic enzymes, DAF-9 appears to act at
a key position in the pathway where it generates nu-
clear-receptor ligands. The selectivity of DAF-9 for 3-
keto-sterols implies that enzymes analogous to mamma-
lian 3b-hydroxysteroid dehydrogenases (3b-HSD), which
are essential for production of all active steroid hormones
in vertebrates, as well as the short-chain oxidoreductases
(SCOR), participate upstream of the production of these
DAF-9 substrates. Interestingly, a search of the C. elegans
genome reveals four genes (Y6B3B.11, ZC8.1, C32D5.12,
ZC449.6) homologous to human 3b-HSD family members
and 84 genes homologous to the SCOR family. Already,
a recent finding by our group suggests a unique Rieske-
like oxygenase, daf-36, is involved in the conversion of
cholesterol into 7-dehydrocholesterol, a potential precur-
sor of D7-dafachronic acid (Rottiers et al., 2006). Finally,
we note that DAF-9 is a functional ortholog of mammalian
CYP27A1, a cytochrome P450 that yields sterol-derived
ligands for the bile acid receptor, FXR (Russell, 2003).
These findings support the notion that oxidation of sterols
to generate signaling molecules coincided with nuclear-
receptor ligand binding and was an acquired trait that oc-
curred throughout evolution.
Multiple Endocrine Networks Control Hormone
the identification of a network of genes that couples
environmental signals to the selection of alternative devel-
opmental pathways of dauer arrest or reproduction. In fa-
vorable environments, these networks utilize peptide hor-
mones such as insulin and TGFb to positively influence
hormone production in C. elegans. Although the exact
mechanisms have not been elucidated, our results dem-
onstrate that this regulation may lie atthe level of hormone
production as D4-dafachronic acid potently rescued the
dauer arrest phenotypes of mutations in daf-2 and daf-7.
Surprisingly, strong daf-2 mutants were rescued from da-
uer but could not fully execute reproductive development.
This result implies an additional requirement for insulin
signaling in reproductive development that may lie parallel
to hormone action. Finally, we show that D4-dafachronic
acid rescues the Daf-c phenotype of the ncr-1;ncr-2 mu-
tant, confirming previous findings that the products of
these genes likely regulate hormone signaling upstream
of daf-9 at the level of substrate availability (Li et al.,
2004). It will be interesting to see whether similar networks
link dietary cues to reproduction in vertebrates.
As demonstrated by this work, a distinct advantage to
working in C. elegans as a model system is the ability to
to elucidate physiologic pathways. Given the previous
success of nuclear receptor pharmacology in vertebrates,
this work also raises the possibility of targeting nuclear
receptors in parasitic nematodes as a strategy for control-
ling their growth. The discovery of dafachronic acids as
1220 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
nematode hormones and the adoption of the orphan nu-
clear receptor DAF-12 set an important precedent for fur-
C. elegans orphan receptors. For now, that’s one down
and 283 to go..
4-cholesten-3-one, lathosterol, 20S-hydroxycholesterol, 22S-hydroxy-
hydroxycholesterol, 3-keto-lithocholic acid, 7-keto-lithocholic acid,
(25S),26-hydroxycholesterol, and (25R),26-hydroxycholesterol were
purchased from Research Plus. Deuterated chenodeoxycholic acid
(CDCA-d4) was from C/D/N Isotopes. Unless otherwise noted, all other
reagents were purchased from Sigma.
3-Keto-D4-oxysterol derivatives were generated with cholesterol oxi-
dase and catalase as described (Zhang et al., 2001). Lathosterone
and lophenone were generated by reacting 2.5 molar equivalents of
Dess-Martin reagent with one equivalent of each sterol at 25C, and
then purified by silica chromatography with 95:5 hexane:ethylacetate.
To oxidize the alcohols into C-26 acids, Jones reagent (0.14 ml, 0.15
mmol) was added dropwise to a stirred solution of each alcohol (12
mg, 0.03 mmol) in acetone (4 ml) at 0ºC. After stirring for 1 hr, the reac-
tion was quenched with isopropanol, and the product was extracted
with diethyl ether. The organic phase was washed with saturated
NaHCO3,dried over solid Na2SO4,filtered, and concentrated in vacuo.
Crude extracts were chromatographed on silica gel, and the product
(10 mg, 90% yield) eluted with 40% ethyl acetate in hexane. All struc-
tures were confirmed by MS, UV spectra, and13C- (data not shown)
and1H-NMR (Table S1).
Nematode and Bacterial Strains
Worms were grown on NGM agar with OP50 bacteria at 20ºC unless
noted otherwise (Brenner, 1974). Strains used were as follows: daf-
9(dh6) dhEx24 (containing the cosmid T13C5 and pTG96 [sur-
5::gfp]), daf-9(rh50), daf-12(rh273), and daf-12(rh61); daf-2(e1368),
daf-2(e1370), daf-7(m62), and N2 (from the CGC); ncr-1(nr2023)ncr-
2(nr2022) (from J.H. Thomas); and daf-9(e1406);daf-12(m20) (from D.
Mammalian expression plasmids wereclonedintoCMX vectors (Ume-
sono et al., 1991; Willy et al., 1995). NHR-8 (NM_171382, wormbase
F33D4.1a), NHR-23 (C01H6.5b), DAF-9 (NM_171699, wormbase
T13C5.1b), and DIN-1S (wormbase F07A11.6d) were obtained by
RT-PCR from mixed stage or L2-/L3-staged animals. Other cDNAs
used were DAF-12 (from K. Yamamoto); CMV-hCYP27A1, CMV-
mCYP27A1, CMV-adrenodoxin (from D. Russell); human P450 oxido-
reductase (Open Biosystems). GAL4 DNA binding domain fusions
were generated with amino acid sequences 184–754 aa (DAF-12),
92–561 aa (NHR-8), 78–361 aa (NHR-23), 2–567 aa (DIN-1S). The
VP16 activation domain was fused to residues 2–754 of DAF-12 to
make VP16-DAF-12. DAF-12 mutants were generated by site-directed
mutagenesis. The 4.2 genomic fragment from pODLO_82 (from K. Ya-
mamoto) was inserted into the reporter plasmid tk-luc to make lit-1K-
tk-luc. DAF-9 and hOR baculovirus expression plasmids were created
using the pFastBac Dual system (Invitrogen).
Cell Culture and Cotransfection Assay
Cotransfections in HEK293 cells were performed in 96-well plates as
described (Makishima et al., 1999) using 50 ng of luciferase reporter,
20 ng CMX-b-galactosidase reporter, 15 ng of CMX receptor expres-
sion plasmid, and control plasmid to maintain 150 ng/well. Candidate
ligandswere added at 4000-fold dilution 8 hr posttransfection. Lucifer-
ase activities were normalized to the b-galactosidase control. Data
represent the mean ± SD of triplicate assays.
Preparation of DAF-9 and Control Microsomes from Sf9 Cells
culovirus (MOI = 2–4) in medium containing 0.5mg/ml hemin chloride,
vested 60 hr post-infection and microsomes prepared as described
(Hood et al., 1996).
DAF-9 Microsomal Incubations
Microsomes containing DAF-9/hOR or hOR alone (as a control) were
generated from Sf9 cells, thawed on ice, and brought to 0.5 mg/ml in
0.1 M potassium phosphate buffer containing a NADPH regenerating
system (50 U/ml DL-isocitrate dehydrogenase, 0.1 M isocitrate and
0.1 M MgCl2). Substrates were added at 100 mM in 0.5 ml total volume,
preincubated 3 min at 37C, and then reacted with 1 mM NADPH for
16 hr. Reactions were processed by extracting twice with 2 ml methyl-
tert-butyl-ether, combining the top layers and drying under nitrogen. In
some experiments, 0.5 mg of 1,4-cholestadiene-3-one was added as
an internal standard for the extraction.
Microsomal extracts were resuspended in 50 ml methanol, mixed with
5? concentration of HB101 bacterial paste, vacuum dried, resus-
pended in 100 ml 5? concentrated HB101, and plated on 3 cm plates
containing 4 ml NG agar. For rescue, ?200 embryos from a 4–8 hr egg
laying were transferred onto the dried bacterial lawn. Mixtures of daf-
9(+),gfp(+) and daf-9(?),gfp(?) embryos were placed on agar plates
containingamixture ofbacteria andextracts from either DAF-9 orcon-
trol microsomal reactions. GFP expressing worms were removed after
48 hr and the remaining daf-9(?),gfp(?) animals were scored for dauer
arrest 24 hr later. For rescue experiments using pure steroids, 10 ml of
compounds were mixed with 5? (90 ml) concentrated OP50 bacteria
and plated. Final concentrations refer to the amount plated in agar
(3–4 ml/plate). Strains tested were grown reproductively onto regular
NG agar for two generations at 20C.
C. elegans Lipid Extracts
Worms were grown on twenty 10 cm NGM plates seeded with HB101
bacteria. Gravid adults were bleached and the resulting embryos incu-
bated in 2.8 liter Fernbach flasks containing 100–350 ml S medium
supplemented with 5 mg/ml Nystatin, 50 mg/ml streptomycin sulfate
overnight to allow synchronization of L1s (Stiernagel, 1999). Two to
three successive rounds of growth (with 1%–2% HB101) and lysis of
gravid adults were performed until ?20-100 million synchronized L1
larvae were obtained. In Figures 7A and 7B, growth of wild-type and
daf-9;daf-12 was performed in Fernbach flasks at 22.5C for 48 hr until
just prior to dorsal turn of distal tip cells. In Figures 7C–7F, final growth
to the L3/L4 stage was performed in a 15 l New Brunswick BiofloIV fer-
25% O2saturation). Worms were harvested and bacteria and debris
were removed by sucrose flotation, then frozen in liquid nitrogen and
stored at ?80C. Thawed worms were lyophilized for measurement of
dry weight, resuspended in 0.1M NaCl, and homogenized using an
Emulsi-flex C-5 homogenizer (Avestin). Total lipids (plus 1 mg CDCA-
d4/107worms) were extracted with 2:1 chloroform:methanol. The re-
sulting chloroform layer was back-extracted with two-thirds volume
of water. The final chloroform layer was dried with Na2SO4, filtered
through Whatman filter paper and concentrated in vacuo. The result-
ing extract (?100 mg/107worms) was resuspended in chloroform,
adsorbed to a silica column, and lipids eluted in three fractions
with 100 ml chloroform, 200 ml 9:1 acetone:methanol, and 100 ml
methanol/100 mg extract (Figure 7C). The 9:1 acetone:methanol
extract was further fractionated by silica chromatography using
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1221
chloroform and increasing concentrations of methanol to 100%. Frac-
tions were dried under nitrogen and tested for DAF-12 activation.
single quadropole instrument (Agilent Technologies) with API-ES in
both positive and negative ion modes. Samples were dissolved in
methanol and loaded onto a precolumn (Zorbax C8, 4.6 ? 12.5 mm,
5 mm, Agilent) at 4 ml/min for 1 min with 30:70 methanol/water, both
containing 5mMNH4Ac,andthenback flushedontotheanalyticalcol-
phase consisted of methanol (A) and methanol/acetonitrile/water
(60:20:20) (B), both containing 5 mM (NH4Ac). The following gradient
was run for a total of 20 min: 0–6.5 min, 75% to 100% (A); 6.5–18
min, 100% (A); 18.1–20 min, 75% (A). MS parameters were as follows:
gas temperature 350ºC, nebulizer pressure 30 psig, drying gas (nitro-
gen) 12 l/min, VCap (positive and negative) 4000V, fragmentor voltage
150V (positive ions) or 200V (negative ions). For experiments in scan
mode, mass ranges between m/z 250–500 were used. Using SIM (in
positive-ion mode), signals for [M + H]+ions were observed for 4-cho-
lesten-3-one (m/z 385, retention time [RT] 12.5 min), lathosterone (m/z
384, RT 14.0 min), 1,4-cholestadien-3-one (m/z 383, RT 10.2 min),
(25R/S),26-hydroxy-4-cholesten-3-one (m/z 401, RT 5.7 min), (25R/
S),26-3-keto-4-cholestenoic acid (m/z 425, RT 4.0 min). SIM LC/MS
in negative-ion mode gave signals for [M ? H]?ions of (25R/S),26-3-
keto-4-cholestenoic acid (m/z 413, RT 4.0 min) and CDCA-d4(m/z
395, RT 3.9 min). Positive and negative ions were monitored simulta-
neously in SIM mode. Separation of 4-cholesten-3-one oxysterols
was achieved as described (Uomori et al., 1987).
Calculation of Endogenous DAF-12 Ligand Concentration
The efficiency of lipid extraction using the CDCA-d4internal standard
was 71% for wild-type and 85% for daf-9(e1406);daf-12(m20) worms.
Crude lipid extracts (792 mg from 50 ? 106wild-type worms, 230 mg
from 19 ? 106daf-9 null worms) were suspended in methanol at 100
mg/ml, filtered through a PVDF membrane, diluted 10-fold and in-
tration of the m/z 413 peak was 81 ng/ml (208 nM/worm) based on an
external calibration curve using 3-keto-4-cholestenoic acid and an L4-
stage worm volume of 1.5 nl.
DAF-12 ligand binding domain (aa 507–753) was expressed in
BL21(DE3) cells as a 6? His-GST fusion protein using pET24a (Nova-
gen). Ligand binding was determined by AlphaScreen assays from
Perkin-Elmer (Xu et al., 2002) with 40 nM receptor and 40 nM of biotin-
ylated SRC1-4 (QKPTSGPQTPQAQQKSLLQQLLTE) peptide in the
presence of 5 mg/ml donor and acceptor beads in a buffer containing
50 mM MOPS, 50 mM NaF, 50 mM CHAPS, and 0.1 mg/ml bovine
serum albumin at pH 7.4. EC50binding values were determined from
nonlinear least square fit of the data based on an average of three
Supplemental Data includefive figures andone table andcanbefound
with this article online at http://www.cell.com/cgi/content/full/124/6/
We thank Jim McKay, Leon Avery, Sylvain Lebreton, Jian Chen, Joon-
Cheol Kwon, Rebecca Lehotzky, and Daniel Schmidt for technical as-
sistance; Keith Yamamoto and David Russell for plasmids; Jeffrey
McDonald from the Lipid Maps Project (GM069338); and Donald
Riddle and J.H. Thomas for worm strains. This work was supported
by the Howard Hughes Medical Institute (D.J.M.); Robert A. Welch
Foundation grants I-1275 (D.J.M.) and I-1493 (R.J.A.); National Insti-
tutes of Health grants GM07062 (D.L.M.), DK62434 and AG027498
(D.J.M. and A.A.), DK0716620 (H.E.X.), and DK59942 (R.J.A.); Jay
and Betty Van Andel Foundation (H.E.X.); Department of Defense
W81XWH0510043 (H.E.X.); and Glenn/AFAR Breakthroughs in Geron-
tology (A.A). D.J.M. is an investigator and C.L.C is an associate of the
Howard Hughes Medical Institute.
Received: December 6, 2005
Revised: January 24, 2006
Accepted: January 31, 2006
Published online: March 9, 2006
Albert, P.S., and Riddle, D.L. (1988). Mutants of Caenorhabditis ele-
gans that form dauer-like larvae. Dev. Biol. 126, 270–293.
Antebi, A., Culotti, J.G., and Hedgecock, E.M. (1998). daf-12 regulates
developmental age and the dauer alternative in Caenorhabditis ele-
gans. Development 125, 1191–1205.
Antebi, A., Yeh, W.H., Tait, D., Hedgecock, E.M., and Riddle, D.L.
(2000).daf-12encodes anuclearreceptor thatregulatesthedauerdia-
pause and developmental age in C. elegans. Genes Dev. 14, 1512–
Brenner, S. (1974). The genetics of Caenorhabditis elegans. Genetics
Cali, J.J., and Russell, D.W. (1991). Characterization of human sterol
27-hydroxylase. A mitochondrial cytochrome P-450 that catalyzes
multiple oxidation reaction in bile acid biosynthesis. J. Biol. Chem.
Chitwood, D.J. (1999). Biochemistry and function of nematode ste-
roids. Crit. Rev. Biochem. Mol. Biol. 34, 273–284.
Chitwood, D.J., Lusby, W.R., Lozano, R., Thompson, M.J., and Svo-
boda, J.A. (1983). Novel nuclear methylation of sterols by the nema-
tode Caenorhabditis elegans. Steroids 42, 311–319.
Gems, D., Sutton, A.J., Sundermeyer, M.L., Albert, P.S., King, K.V.,
Edgley, M.L., Larsen, P.L., and Riddle, D.L. (1998). Two pleiotropic
classes of daf-2 mutation affect larval arrest, adult behavior, reproduc-
tion and longevity in Caenorhabditis elegans. Genetics 150, 129–155.
9/cytochromeP450 regulateC.elegans dauer diapause in response to
environmental cues. Development 131, 1765–1776.
Gerisch, B., Weitzel, C., Kober-Eisermann, C., Rottiers, V., and Antebi,
A. (2001). A hormonal signaling pathway influencing C. elegans metab-
olism, reproductive development, and life span. Dev. Cell 1, 841–851.
Gill, M.S., Held, J.M., Fisher, A.L., Gibson, B.W., and Lithgow, G.J.
(2004). Lipophilic regulator of a developmental switch in Caenorhabdi-
tis elegans. Aging Cell 3, 413–421.
Gissendanner, C.R., Crossgrove, K., Kraus, K.A., Maina, C.V., and
Sluder, A.E. (2004). Expression and function of conserved nuclear re-
ceptor genes in Caenorhabditis elegans. Dev. Biol. 266, 399–416.
Hood, S.R., Shah, G., and Jones, P. (1996). Expression of cyto-
chromes P450 in a baculovirus system. In Methods in Molecular Biol-
ogy,I.R.Phillips andE.A. Shephard,eds.(Totowa,NJ:HumanaPress),
Hsin, H., and Kenyon, C. (1999). Signals from the reproductive system
regulate the life span of C. elegans. Nature 399, 362–366.
ment 129, 221–231.
Kimura, K.D., Tissenbaum, H.A., Liu, Y., and Ruvkun, G. (1997). daf-2,
an insulin receptor-like gene that regulates longevity and diapause in
Caenorhabditis elegans. Science 277, 942–946.
1222 Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc.
Larsen, P.L., Albert, P.S., and Riddle, D.L. (1995). Genes that regulate
both development and longevity in Caenorhabditis elegans. Genetics
Li, J., Brown, G., Ailion, M., Lee, S., and Thomas, J.H. (2004). NCR-1
and NCR-2, the C. elegans homologs of the human Niemann-Pick
type C1 disease protein, function upstream of DAF-9 in the dauer for-
mation pathways. Development 131, 5741–5752.
Ludewig, A.H., Kober-Eisermann, C., Weitzel, C., Bethke, A., Neubert,
tor/coregulator complex controls C. elegans lipid metabolism, larval
development, and aging. Genes Dev. 18, 2120–2133.
development by the C. elegans DAF-9 cytochrome P450. Develop-
ment 131, 1777–1786.
Makishima, M., Okamoto, A.Y., Repa, J.J., Tu, H., Learned, R.M., Luk,
A., Hull, M.V., Lustig, K.D., Mangelsdorf, D.J., and Shan, B. (1999).
Identification of a nuclear receptor for bile acids. Science 284, 1362–
Matyash, V., Entchev, E.V., Mende, F., Wilsch-Brauninger, M., Thiele,
C., Schmidt, A.W., Knolker, H.J., Ward, S., and Kurzchalia, T.V. (2004).
Sterol-derived hormone(s) controls entry into diapause in Caenorhab-
ditis elegans by consecutive activation of DAF-12 and DAF-16. PLoS
Biol. 2, e280 10.1371/journal.pbio.0020280.
Merris, M., Wadsworth, W.G., Khamrai, U., Bittman, R., Chitwood,
D.J., and Lenard, J. (2003). Sterol effects and sites of sterol accumu-
lation in Caenorhabditis elegans: developmental requirement for
4alpha-methyl sterols. J. Lipid Res. 44, 172–181.
Mooijaart, S.P., Brandt, B.W., Baldal, E.A., Pijpe, J., Kuningas, M.,
Beekman, M., Zwaan, B.J., Slagboom, P.E., Westendorp, R.G., and
vanHeemst,D. (2005). C.elegans DAF-12,nuclearhormone receptors
and human longevity and disease at old age. Ageing Res. Rev. 4, 351–
Norlin, M., von Bahr, S., Bjorkhem, I., and Wikvall, K. (2003). On the
substrate specificity of human CYP27A1: implications for bile acid
and cholestanol formation. J. Lipid Res. 44, 1515–1522.
teintyrosinephosphatase-like molecule, regulatestheL3/dauerdevel-
opmental decision through hormonal signaling in C. elegans. Develop-
ment 130, 3237–3248.
Ren, P., Lim, C.S., Johnsen, R., Albert, P.S., Pilgrim, D., and Riddle,
D.L. (1996). Control of C. elegans larval development by neuronal ex-
pression of a TGF-beta homolog. Science 274, 1389–1391.
Riddle, D.L., and Albert, P.S. (1997). Genetic and environmental regu-
lation of dauer larva development. In C. elegans II, D.L. Riddle, B.
Meyer, J. Priess, and T. Blumenthal, eds. (Colds Spring Harbor, NY:
Cold Spring Harbor Laboratory Press), pp. 739–768.
Rottiers, V., Motola, D.L., Gerisch, B., Cummins, C.L., Nishiwaki, K.,
Mangelsdorf, D.J., and Antebi, A. (2006). Hormonal control of C. ele-
gans dauer formation and life span by a Rieske-like oxygenase. Dev.
Cell, in press.
Russell, D.W. (2003). The enzymes, regulation, and genetics of bile
acid synthesis. Annu. Rev. Biochem. 72, 137–174.
Schackwitz, W.S., Inoue, T., and Thomas, J.H. (1996). Chemosensory
neurons function in parallel to mediate a pheromone response in
C. elegans. Neuron 17, 719–728.
Shostak, Y., Van Gilst, M.R., Antebi, A., and Yamamoto, K.R. (2004).
Identification of C. elegans DAF-12-binding sites, response elements,
and target genes. Genes Dev. 18, 2529–2544.
Sluder, A.E., and Maina, C.V. (2001). Nuclear receptors in nematodes:
themes and variations. Trends Genet. 17, 206–213.
Stiernagel, T. (1999). Maintenance of C. elegans. In C. elegans: A.
Practical Approach, I.A. Hope, ed. (New York: Oxford University
Press), pp. 51–67.
Thomas, J.H., Birnby, D.A., and Vowels, J.J. (1993). Evidence for par-
allel processing of sensory information controlling dauer formation in
Caenorhabditis elegans. Genetics 134, 1105–1117.
Umesono, K., Murakami, K.K., Thompson, C.C., and Evans, R.M.
(1991). Direct repeats as selective response elements for the thyroid
hormone, retinoic acid, and vitamin D3 receptors. Cell 65, 1255–1266.
Uomori, A., Seo, S., Sato, T., Yoshimura, Y., and Takeda, K. (1987).
olution of (25R)- and (25S)-26-Hydroxycholesterol. J. Chem. Soc.
Perkin Trans. I 1, 1713–1718.
Van Gilst, M.R., Hadjivassiliou, H., Jolly, A., and Yamamoto, K.R.
(2005a). Nuclear hormone receptor NHR-49 controls fat consumption
and fatty acid composition in C. elegans. PLoS Biol. 3, e53 10.1371/
Van Gilst, M.R., Hadjivassiliou, H., and Yamamoto, K.R. (2005b). From
the cover: a Caenorhabditis elegans nutrient response system partially
dependent on nuclear receptor NHR-49. Proc. Natl. Acad. Sci. USA
Vowels, J.J., and Thomas, J.H. (1992). Genetic analysis of chemosen-
sory control of dauer formation in Caenorhabditis elegans. Genetics
Willy, P.J., Umesono, K., Ong, E.S., Evans, R.M., Heyman, R.A., and
retinoid response pathway. Genes Dev. 9, 1033–1045.
Xu, H.E., Stanley, T.B., Montana, V.G., Lambert, M.H., Shearer, B.G.,
Cobb, J.E., McKee, D.D., Galardi, C.M., Plunket, K.D., Nolte, R.T.,
et al. (2002). Structural basis for antagonist-mediated recruitment of
nuclear co-repressors by PPARalpha. Nature 415, 813–817.
ketone derivatives by HPLC and response to physiological perturba-
tions. J. Lipid Res. 42, 649–658.
Cell 124, 1209–1223, March 24, 2006 ª2006 Elsevier Inc. 1223